WO2005081539A1 - 2次元信号の符号化/復号方法および装置 - Google Patents
2次元信号の符号化/復号方法および装置 Download PDFInfo
- Publication number
- WO2005081539A1 WO2005081539A1 PCT/JP2004/019604 JP2004019604W WO2005081539A1 WO 2005081539 A1 WO2005081539 A1 WO 2005081539A1 JP 2004019604 W JP2004019604 W JP 2004019604W WO 2005081539 A1 WO2005081539 A1 WO 2005081539A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- coefficient
- encoding
- coefficients
- map
- region
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/59—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial sub-sampling or interpolation, e.g. alteration of picture size or resolution
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/124—Quantisation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/132—Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/14—Coding unit complexity, e.g. amount of activity or edge presence estimation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/1883—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit relating to sub-band structure, e.g. hierarchical level, directional tree, e.g. low-high [LH], high-low [HL], high-high [HH]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/423—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/63—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/63—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets
- H04N19/64—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets characterised by ordering of coefficients or of bits for transmission
Definitions
- the present invention relates to a two-dimensional signal compression / expansion method, and more particularly to a method and apparatus for encoding a two-dimensional signal using a wavelet transform.
- Patent Document 1 Japanese Patent Application Laid-Open No. 9-1148938
- variable length coding is performed on each of the divided subbands, and codes are output in order from the low frequency subband.
- the merits of dividing into tiles are that the neighboring areas with similar statistical properties of the coefficients are collected, and that the coding efficiency can be improved by applying the optimal coding to each area.
- parallel processing and decoding can be realized.
- the division into tiles is usually determined to be about 64 ⁇ 64 or more, since over-division reduces the code efficiency.
- the method of performing wavelet transform after tile division has a problem that discontinuous distortion, that is, block noise occurs at the tile boundary.
- Patent Document 2 Japanese Patent Laid-Open No. 2000-316152 divides each subband after wavelet transform into blocks (8 ⁇ 8), and And perform encoding.
- the advantage of dividing into blocks after wavelet transform is that block noise does not occur.
- Patent Document 3 Japanese Patent Application Laid-Open No. H11-298897 discloses an image data transmission method in which the size of a displayed image increases as the resolution increases.
- Patent Document 4 Japanese Patent Application Publication No. 2003-504921 discloses a method of dividing an image and rendering each sub-image at a different wavelet decoding rate. Specifically, utilizing the fact that the center of the screen is the area that is most closely watched by the user, the detail is drawn first in the sub-image at the center, and the detail is drawn gradually as the distance from the center increases. Is done. Even with such a display function, the cognitive waiting time of the image viewer can be reduced.
- the CPU encodes a coefficient with respect to a memory address where each coefficient value is stored.
- the address force S where the coefficient is stored is located at a position distant from the scan line, and extends over many rows. For example, if the 2LH sub-band is divided into blocks (8 ⁇ 8), the memory access to 8 scan lines is performed to expand 2LH. For this reason, the memory access destinations are dispersed, and the efficiency of the cache memory is reduced. Since a general CPU has a cache memory, the local memory access can be performed at high speed. If the memory access destinations are distributed as described above, access to the real memory occurs, and the processing speed is reduced. Significantly reduced.
- the above-described conventional method has a problem that a large intermediate memory is required.
- the wavelet transform of a certain tile or block is started, and at the time of decoding, until the encoding of the tile or block is completed, at the time of decoding, the decoding of the coefficient of a certain tile or block is started and then the reverse.
- the LH, HL, and HH subbands need to be stored in memory until the wavelet transform is completed.
- an object of the present invention is to provide a code Z decoding method and apparatus capable of obtaining a sufficient speed improvement result even when performing wavelet coding / decoding using a sequential processing type CPU. It is in.
- Still another object of the present invention is to provide an encoding / decoding method and a coding / decoding method capable of reducing the capacity of an intermediate memory required for temporarily storing subbands in wavelet encoding / decoding processing. It is to provide a device.
- a code encoding apparatus includes: a wavelet transform unit that divides a two-dimensional signal into a plurality of frequency-domain sub-bands; and a plurality of sub-bands belonging to the same hierarchical wavelet decomposition level. It is characterized by comprising coefficient extraction means for extracting a set of coefficients belonging to the same position spatially from the band by a fixed number, and coefficient coding means for coding the extracted set of coefficients.
- An encoding device includes: an element extraction unit configured to sequentially extract 2mx2 (m is an integer: m ⁇ l) spatially adjacent elements from a two-dimensional signal; and Wavelet transform means for dividing the coefficient set into a plurality of subbands, coefficient coding means for coding the coefficient set, and a code for rearranging and outputting the coded coefficient set in order from the subband on the low resolution side And output means.
- an element extraction unit configured to sequentially extract 2mx2 (m is an integer: m ⁇ l) spatially adjacent elements from a two-dimensional signal
- Wavelet transform means for dividing the coefficient set into a plurality of subbands, coefficient coding means for coding the coefficient set, and a code for rearranging and outputting the coded coefficient set in order from the subband on the low resolution side And output means.
- a decoding device receives a code system 1J obtained by encoding coefficients of a plurality of subbands obtained by wavelet transform, and generates an original two-dimensional signal.
- An initial coefficient decoding means for decoding the coefficient of the lowest frequency subband from a code sequence corresponding to the lowest frequency subband, and a code sequence subsequent to the lowest frequency subband code sequence, belonging to a wavelet decomposition level of the same hierarchy.
- an image is divided into sub-bands by a wavelet transform based on the Haar function, and the entire region of the lowest frequency LL sub-band is encoded.
- the LH, HL, and HH subband coefficients belonging to are encoded for each coefficient at the same spatial position.
- the decoding side first expands the lowest frequency LL subband, then decodes one set of spatially identical LH, HL, and HH coefficients in each wavelet decomposition level subband.
- the inverse wavelet transform is immediately performed by using the coefficient value of, and the LL coefficient value of the next wavelet decomposition level is obtained.
- a general-purpose sequential processing type CPU is used to sequentially read and encode coefficients respectively corresponding to the same image position from the LH, HL, and HH subbands.
- the number of stores and loads to the real memory can be reduced and high-speed processing can be performed.Furthermore, since the coefficients at the same position of the LH, HL and HH subbands are simultaneously encoded sequentially, the LH, HL and HH subband There is no need to keep the data until the encoding is completed.
- a high-speed register or cache memory of a CPU can be used as a coefficient memory for storing coefficients to be coded, and the coding speed can be further improved.
- each time a set of LH, HL, and HH subband coefficients are decoded their coefficient values are immediately subjected to an inverse wavelet operation on a register, and access to the real memory can be reduced.
- reading of the LL subband is performed continuously on the same scan line, and writing of the inverse wavelet operation result is also performed continuously on only two scan lines. Therefore, the efficiency of the cache memory is improved. Furthermore, the amount of intermediate memory required to temporarily store subbands can be reduced.
- FIG. 1 is a block diagram showing a schematic configuration of an encoding device according to a first embodiment of the present invention.
- FIG. 2A shows each sub-band nLL, nLH, nHL, nH in the sub-band memory 102.
- FIG. 3 is a schematic diagram illustrating an example of an H coefficient array.
- FIG. 2B is a schematic diagram showing an example of a code array in the output code memory 104.
- FIG. 3 is a flowchart showing encoding control by the encoding device according to the first embodiment of the present invention.
- FIG. 4 is a block diagram showing a schematic configuration of a decoding device according to a first embodiment of the present invention.
- FIG. 5 is a flowchart showing decoding control of the decoding device according to the first embodiment of the present invention.
- FIG. 6 is a block diagram showing a schematic configuration of a coding apparatus according to a second embodiment of the present invention.
- FIG. 7 is a flowchart showing encoding control performed by the encoding device according to the first embodiment of the present invention.
- FIG. 8 is a diagram schematically showing encoding and decoding operations according to the first and second embodiments of the present invention.
- FIG. 9 is a block diagram showing a schematic configuration of an encoding device according to a third embodiment of the present invention.
- FIG. 10 is a flowchart showing encoding control by an encoding device according to a third embodiment of the present invention. It is one chart.
- FIG. 11 is a diagram schematically showing encoding and decoding operations according to a third embodiment of the present invention.
- FIG. 12 is a block diagram showing a schematic configuration of a coding apparatus according to a fourth embodiment of the present invention.
- FIG. 13 is a flowchart showing encoding control by the encoding device according to the fourth embodiment of the present invention.
- FIG. 14 is a diagram schematically showing a console screen for explaining a moving area detection and a thinning-out operation in a fourth embodiment of the present invention.
- FIG. 15 is a block diagram showing a schematic configuration of a coding apparatus according to a fifth embodiment of the present invention.
- FIG. 16 is a flowchart showing encoding control according to a fifth embodiment of the present invention.
- FIG. 17 is a diagram schematically showing encoding and decoding operations according to a fifth embodiment of the present invention.
- FIG. 18 is a block diagram showing a schematic configuration of a code coder according to a sixth embodiment of the present invention.
- FIG. 19 is a flowchart showing the encoding control according to the sixth embodiment of the present invention.
- FIG. 20 is a block diagram showing a schematic configuration of an encoding device according to a seventh embodiment of the present invention.
- FIG. 21 is a flowchart showing the encoding control according to the seventh embodiment of the present invention.
- FIG. 22 is a diagram schematically showing a console screen for explaining a moving region detection and a coefficient quantization operation in the seventh embodiment of the present invention.
- FIG. 23 is a block diagram of a bidirectional communication terminal incorporating a coding and decoding device according to an eighth embodiment of the present invention.
- FIG. 24 is a block diagram showing a schematic configuration of a code coder according to a ninth embodiment of the present invention.
- FIG. 25 is a block diagram showing a schematic configuration of a decoding device according to a ninth embodiment of the present invention.
- FIG. 1 is a block diagram illustrating a schematic configuration of a code apparatus according to a first embodiment of the present invention.
- the encoding device inputs the image data 10 from an image input device (not shown), and outputs a code sequence by wavelet encoding described later.
- the encoding device mainly includes a data processing device 101, a sub-band memory 102, a coefficient memory 103, and an output code memory 104.
- the generated code sequence is output to a code output device 105.
- the data processing device 101 is, for example, a program control processor such as a general CPU (Central Processing Unit), and reads a code group from a program memory (not shown) and executes it to perform wavelet transform (here, The two-dimensional Haar wavelet transform) process (wavelet transform means 101a), the initial coefficient coding process (initial coefficient coding means 101b), the same coordinate coefficient extraction process (same coordinate coefficient extraction means 101c), the coefficient code A conversion process (coefficient encoding means 101d) and a code output process (code output means 101e) can be generated (the details will be described later).
- wavelet transform here, The two-dimensional Haar wavelet transform
- wavelet transform means 101a The two-dimensional Haar wavelet transform
- the initial coefficient coding process initial coefficient coding process
- the same coordinate coefficient extraction process space coordinate coefficient extraction means 101c
- the coefficient code A conversion process coefficient encoding means 101d
- code output means 101e code output means
- the sub-band memory 102, coefficient memory 103, and output code memory 104 can be configured by separate memories, for example, by allocating each space to one semiconductor memory. It is OK to provide a space for storing the input image data 10 in this semiconductor memory. As described later, in the present invention, a register cache memory of a CPU can be used as the coefficient memory 103.
- the subband memory 102 stores N (integer not less than l) hierarchical subband data (nLL, nLH, nHL, nHH) (l ⁇ n ⁇ N).
- N integer not less than l
- the subbands nLL, nLH, nHL, and nHH of the layer n are composed of a coefficient row s IJ of p X q (p, q are integers) corresponding to the spatial coordinates (x, y) of the image.
- the subband data nLL, nLH, nHL, and nHH of each layer are all two-dimensional signals.
- the coefficient memory 103 stores the subbands nLH, nHL, and nHH of one layer that are sequentially extracted from the subband memory 102 by the same coordinate coefficient extraction process of the data processing device 101.
- a set of coefficients ⁇ LH (i, j), HL (i, j), HH (i, j) ⁇ of the same spatial coordinates (i, j) is stored.
- the order in which the coefficient sets are extracted follows the scan line direction from the upper left edge to the lower right edge of the image. However, two or more constant coefficient sets can be sequentially extracted.
- one or more coefficient sets adjacent to ⁇ LH (i, j), HL (i, j), HH (i, j) ⁇ may be extracted and stored in the coefficient memory 103.
- the cache memory of the CPU is used as the coefficient memory 103, a plurality of sets of coefficients can be stored if the cache memory has room.
- the output code memory 104 has an NLL sub-band code memory 104.1 and a (nLH, nHL, nHH) sub-band code memory 104.2.
- the NLL subband code memory 104.1 stores a subband code NLLc obtained by coding the subband NLL of the layer N by the initial coefficient coding process of the data processing device 101.
- the subband code memory 104.2 is a code (nLH, nHLH) obtained by coding the set of the same coordinate coefficients sequentially read out to the coefficient memory 103 by a coefficient coding process.
- NHH stores c
- HH (i, j) ⁇ is stored and encoded.
- FIG. 2A is a schematic diagram showing an example of a coefficient array of each subband nLL, nLH, nHL, nHH in the subband memory 102
- FIG. 2B is a schematic diagram showing an example of a code array in the output code memory 104. It is.
- the coefficient array of subband nLL of layer n obtained by two-dimensional Haar wavelet transform is nL (0, 0), nL (l, 0),..., NL (p, q), subband nLH coefficient nn
- the array is nA (0, 0), nA (l, 0), ⁇ , nA (p, q), and the coefficient array for subband nHL is nB (n n
- nHH nC (0, 0), nC n n
- the NLL subband code memory 104.1 stores the above-described coefficients NL (0, 0), NL (1, 0),..., NL (p, q) of the subband NLL as initial codes. Sign by the conversion process NLLc (0, 0), NLLc (l, 0), ⁇ , NLLc (p, q) are stored. This
- the difference between adjacent coefficients is encoded in the encoding method of the initial encoding process.
- a DPCM (Diiferential Pulse Code Modulation) Huffman code or the like can be used.
- the (nLH, nHL, nHH) subband code memory 104.2 stores one or more sets (nLH, nHL, nHH) of the same coordinate coefficients sequentially read out to the coefficient memory 103 as coefficient codes.
- the codes (nLHc, nHLc, nHHc) that have been set one by one by the conversion process are stored.
- a run-length Huffman code or an LZW (Lempe Ziv-Welch) code using a run length of zero value and a value of a non-zero coefficient can be used as an encoding method of the coefficient encoding process.
- nA (0,0), nB (0,0) having the same coordinates (0,0) of subbands nLH, nHL, nHH of hierarchy n ) And nC (0,0) are collectively read out to the coefficient memory 103, and nAc (0,0), nBc (0,0) and nCc (0,0) encoded by the coefficient encoding process are obtained. It is stored in the subband code memory 104.2.
- the coefficients ⁇ (1, 0), ⁇ (1, 0) and nC (l, 0) having the same coordinates (1, 0) are read together into the coefficient memory 103, and are encoded by the coefficient encoding process.
- the nAc (l, 0), nBc (1,0) and nCc (l, 0) that have been mapped are stored in the subband code memory 104.2.
- all coefficients up to the coordinates (p, q) are encoded and stored in the subband code memory 104.2.
- a coefficient nA (0, 0) having the same coordinates (0, 0) and (1, 0) of subbands nLH, nH L, and nHH of layer n ), NB (0,0), nC (0,0) and coefficients nA (l, 0), nB (l, 0), nC (l, 0) are collectively read out to the coefficient memory 103, and similarly, Encoding is performed and the codes nAc (0,0), nBc (0,0) and nCc (0,0) and nAc (l, 0), nBc (l, 0), nCc (l, 0) It may be stored in the subband code memory 104.2.
- Lc and nHHc are output to the code output device 105 by a code output process.
- the code output device 105 is, for example, a drive that records information on an optical recording medium such as a DVD, or a network interface connected to a network.
- FIG. 3 is a flowchart showing the encoding control in the present embodiment.
- image data 10 is input from an image input device such as a camera or a scanner, and stored in an image memory (step S 11).
- the wavelet transform process on the data processing device 101 executes N-stage two-dimensional Haar wavelet transform on the image data 10 read out from the image memory, and outputs the transform result as described above. Is stored in the sub-band memory 102 (step S12).
- the initial coefficient coding process reads the initial coefficient NLL from the subband memory of the layer N, and uses the run-length Huffman code or the LZW code to perform the variable-length coding.
- the codes NLLc (0, 0), NLLc (l, 0), ..., NLLc (p, q) are stored in the NLL sub-band code memory 104.1 of the output code memory 104.
- the next encoding target is the LH, HL, and HH subbands.
- the variable n indicating the resolution level (hierarchy) is set to N
- the same coordinate coefficient extraction process is nLH, nHL
- the same coordinate coefficient extraction process is based on the same set of coefficients (nA (i, j), nB) of the same spatial coordinates (i, j) in the subbands nLH, nHL, nHH of the hierarchy n from the subband memory 102.
- (i, j), nC (i, j) ⁇ is read out and stored in the coefficient memory 103 (step S15).
- a set of two or more constant coefficients may be read. For example, a fixed number of coefficient sets corresponding to the same spatial coordinates (i, j), (i + 1, j).
- coefficient sets nA (i, j), nB (i, j), and nC (i, j) are obtained from the coefficient memory 103.
- the data is read out, subjected to variable length coding, and stored in the subband code memory 104.2 of the output code memory 104 (step S16).
- the same coordinate coefficient extraction process determines whether or not the coding has been completed for all the coordinates of the layer n (step S17). .
- Step S17 If all the coding in the layer n has not been completed (N ⁇ in step S17), the extraction coordinates (i, j) are shifted by a predetermined number in the scan line direction (step S18). Steps S15 and S18 are repeated until all encoding in n is completed.
- the code output process reads the code string stored in the output code memory 104 and outputs the code. Output to the device 105 (step S22).
- the code is output after the completion of the entire encoding.
- the code may be output every time a fixed amount of codes are accumulated.
- the low-frequency component of each resolution is stored in the image memory and each nLL sub-band memory, but the present invention is not limited to these storage methods. For example, you can take the method of sharing these memories and reducing the amount of memory used.
- FIG. 4 is a block diagram showing a schematic configuration of a decoding device according to the first embodiment of the present invention.
- the decoding device inputs a code string 20 output from the coding device, The original image data is output by bullet decoding.
- the decoding device mainly has a data processing device 201, a sub-band memory 202, a coefficient memory 203, and an image output device 204, and the decoded image data is output to the image output device 204.
- the data processing device 201 is a program control processor such as a CPU, similar to the encoding device.
- the data processing device 201 reads a decoding program group from a program memory (not shown) and executes it to execute a decoding process of initial coefficients (initial coefficient Decoding means 201a), the same coordinate coefficient decoding process (the same coordinate coefficient decoding means 201b), the inverse wavelet transform (here, the inverse two-dimensional Harle wavelet transform) process (the inverse wavelet transform means 201c), and the image output A process (image output means 201d) can be generated (details will be described later).
- the sub-band memory 202 and the coefficient memory 203 can be configured by separate memories, respectively, but can also be configured by allocating respective spaces to one semiconductor memory, for example.
- An input code memory for storing the input code string 20 may be provided in this semiconductor memory.
- a register or a cache memory of the CPU can be used as the coefficient memory 203.
- the sub-band memory 202 sequentially generates (N-1) LL, (N_2) from the decoded NLL sub-band coefficient sequence by a decoding step of the initial coefficient and the same coordinate coefficient and a wavelet inverse transformation step described later. ) LL, ⁇ ⁇ ⁇ Stores a subband coefficient sequence.
- NLL subband coefficient system IJ we generalize including the NLL subband coefficient system IJ and describe it as an nLL subband coefficient sequence (0 ⁇ n ⁇ N).
- the coefficient memory 203 stores a set of subband coefficients ⁇ LH (i, j), HL (i, j) of the same spatial coordinates (i, j) decoded by the same coordinate coefficient decoding process of the data processing device 201. ), HH (i, j) ⁇ .
- two or more constant coefficient sets may be stored. For example, if one or more coefficient sets adjacent to ⁇ LH (i, j), HL (i, j), HH (i, j) ⁇ are coded, multiple coefficient sets are decoded and the coefficient memory Is stored in As described above, when the cache memory of the CPU is used as the coefficient memory 203, a plurality of sets of coefficients can be stored if the cache memory has room.
- the image output device 204 is, for example, a liquid crystal display.
- the image output process of the data processing device 201 is a process of decoding the initial coefficient and the same coordinate coefficient, which will be described later, and the image data decoded by the inverse wavelet transform process. Display in 4.
- FIG. 5 is a flowchart showing the decoding control of the decoding device according to the first embodiment of the present invention.
- the input variable-length code sequence 20 is stored in the input code memory (step S31).
- variable length code of the NLL sub-band is sequentially read from the input code memory and decoded, and the obtained initial coefficients NLL sub-band coefficients NL (0, 0), NL ( 0, 1) ⁇ ⁇ ⁇ ⁇ are stored in the NLL sub-band memory 202 (step S32).
- the next decoding target is the LH, HL, and HH subbands.
- the variable n indicating the resolution level (hierarchy) is set to N
- the same coordinate coefficient decoding process is nLH, nHL, nHH
- the same coordinate coefficient decoding process reads the code strings nAc (i, j), nBc (i, j), nCc (i, j) from the input code memory, and outputs the subband nLH , NHL, nHH, the set of coefficients of the same spatial coordinates (i, j) ⁇ nA (i, j), nB (i, j), nC (i, j) ⁇ is stored in the coefficient memory 203 (step S34). .
- the plurality of sets of coefficients are stored in the coefficient memory 203.
- the inverse wavelet transform process performs the inverse wavelet transform.
- the inverse wavelet transform process calculates the coefficients nA (i, j), nB (i, j), nC (i, j) from the coefficient memory 203 and the corresponding nLL subband coefficients nL (i , j) are read out (step S35). Wavelet inverse transform is performed on the read nL (i, j), nA (i, j), nB (i, j) and nC (i, j), and the result is stored in the subband memory 202. (n-1) The coefficient of the LL subband (n-1) is stored as L (i, j) (step S36). Subsequently, the same coordinate coefficient decoding process determines whether decoding has been completed for all coordinates of the layer n (step S37).
- step S38 If all the decoding in the layer n has not been completed (N in step 37), the coordinates (i, j) are shifted by a predetermined number in the scan line direction (step S38), and in the layer n . Steps S34 to S38 are repeated until all decoding is completed. [0060] When all decoding on the layer n is completed (YES in step S37), it is determined whether or not the current layer n has reached the layer N corresponding to the resolution to be output (step S39).
- n is decremented by 1 (step S39).
- the steps S34, S40, and S301 are repeated until n reaches N while decreasing the layer n by one.
- the output process reads (n_l) LL sub-band data from the sub-band memory 202, outputs the data to the image output device 204, and displays the image (step S302).
- the decoding of the sub-band of the hierarchy n 1, which is the highest resolution level, is completed (YES in step S303), the process ends.
- the decoding operation is performed after all the codes have been input.
- the decoding may be performed every time a fixed amount of codes are input sequentially from the initial codes.
- the input code sequence is stored in the input code memory, and the nLL subband coefficient is stored in the nLL subband memory.
- the present invention is not limited to these storage methods. .
- a method of sharing a part of these memories and reducing the amount of used memory may be adopted.
- the code output by the coefficient encoding process of the encoding device and the code input to the coefficient decoding process of the decoding device are:
- the LH, HL, and HH coefficients of the Y component may be arranged first, then the LH, HL, and HH coefficients of the Cb component, and finally the LH, HL, and HH coefficients of the Cr component.
- SIMD in which the CPU that performs encoding and decoding applies the same operation to multiple data strings
- LH, HL, and HH subband coefficients obtained by performing wavelet transform on image data are stored, and coefficients corresponding to the same position in the image, respectively. are sequentially read at least one set at a time, encoded, and stored in the output code memory.
- a high-speed register or cache memory of the CPU can be used as a coefficient memory for storing coefficients to be coded, and the coding speed can be improved.
- one or more sets of LH, HL, and HH subband coefficients corresponding to the same image position decoded by the coefficient decoding process are stored in registers and the like.
- the wavelet is inversely transformed as soon as it is read into the high-speed coefficient memory. Therefore, in the sequential processing type CPU, the number of times of storing and loading to the real memory can be reduced, and the decoding speed is greatly increased.
- the inverse wavelet transform process continuously reads the LL sub-bands in the order of scan lines, and outputs the operation results on two scan lines of the nLL sub-bands continuously. And speeding up can be achieved.
- FIG. 6 is a block diagram showing a schematic configuration of a code apparatus according to a second embodiment of the present invention.
- the encoding device inputs image data 10 as a representative example of a two-dimensional signal from an image input device (not shown), and outputs a code sequence by wavelet encoding described later.
- the encoding device mainly includes a data processing device 301, an LL sub-band memory 302, a coefficient memory 303, and an output code memory 304.
- the generated code sequence is output to a code output device 305.
- the data processing device 301 is a general program control processor such as a CPU, and reads and executes a group of encoded programs from a program memory (not shown) to execute a pixel extraction process (pixel extraction (element extraction) means. 30 la), wavelet transform (here, 2D Haar wavelet transform) process (wavelet transform means 301b), coefficient encoding process (coefficient encoding means 301c), initial coefficient encoding process (initial coefficient encoding means 301d) , And a code output process (code output means 301e) can be generated (the details will be described later).
- a pixel extraction process pixel extraction (element extraction) means. 30 la
- wavelet transform here, 2D Haar wavelet transform
- coefficient encoding process coefficient encoding means 301c
- initial coefficient encoding process initial coefficient encoding process
- initial coefficient encoding means 301d initial coefficient encoding means 301d
- code output means 301e can be generated (the details will be described later).
- the LL sub-band memory 302, coefficient memory 303, and output code memory 304 are configured by allocating respective spaces to one semiconductor memory, for example, by allocating respective spaces to one semiconductor memory. It is also possible. Note that, as described later, in the present invention, a register or a cache memory of the CPU can be used as the coefficient memory 303.
- the LL sub-band memory 302 stores the input image data 10 as 0LL sub-band data, and is obtained by performing a two-dimensional Haar wavelet transform on the partial area (2mx2) sequentially extracted by the pixel extraction process.
- the nLL subband coefficients (1 ⁇ n ⁇ N) are stored sequentially, and finally the NLL subband coefficients are stored.
- the pixel extraction process sequentially extracts the spatially adjacent 2m ⁇ 2 pixel partial regions from the (n ⁇ 1) LL subband coefficients, and performs a two-dimensional Haar wavelet transform on them. Then, the nLL subband coefficients are obtained.
- the extraction order of the 2mx2 partial area follows the scan line direction from the upper left corner to the lower right corner of the image.
- the coefficient memory 303 stores a set of coefficients ⁇ LH (i, j), HL of the same spatial coordinates (i, j) in one layer of subbands nLH, nHL, nHH obtained by the two-dimensional Haar wavelet transform. (i, j), HH (i, j) ⁇ .
- two or more constant coefficient sets may be stored.
- a plurality of sets of coefficients can be stored if the cache memory has room.
- the output code memory 304 has a (nLH, nHL, nHH) subband code memory 304.1 and an NLL subband code memory 304.2 as an initial output code.
- the sub-band code memory 304.1 is a code (nLH, nHLH) obtained by coding a set of identical coordinate coefficients sequentially read out to the coefficient memory 303 by a coefficient coding process. , NHH) c.
- the NLL subband code memory 304.2 stores the subband code NLLc obtained by coding the NLL subband coefficients in the initial coefficient coding process of the data processing device 301.
- FIG. 7 is a flowchart showing the encoding control in this embodiment.
- image data 10 is input from an image input device such as a camera or a scanner, and stored in the LL sub-band memory 302 as a 0LL sub-band (step S41).
- Step S43 the upper left 2 ⁇ 2 sub-region (HI, H2, H3, H4) is extracted.
- a two-dimensional Haar wavelet transform is performed on the extracted partial area of 2m x 2 pixels (step S44), and the resulting LH coefficient nA (i, j), HL coefficient
- the nB (i, j) and the HH coefficient nC (i, j) are stored in the coefficient memory 303, respectively, and the LL coefficient nL (i, j) is stored in the LL subband memory 302 as an nLL subband (step S45).
- the wavelet transform process calls the coefficient encoding process, and the coefficient memory 3
- the coefficient sets nA (i, j), nB (i, j), and nC (i, j) are read from 03, subjected to variable-length coding, and stored in the subband code memory 304.1 of the output code memory 304. (Step S46).
- the pixel extraction process determines whether or not the encoding has been completed for all the coordinates of the layer n (step S47).
- step S47 If all the coding in the layer n has not been completed (N in step S47), the extraction coordinates (i, j) are shifted by a predetermined number in the scan line direction (step S48). Steps S43 and S48 are repeated until all encoding in n is completed.
- step S49 When the encoding of the subband of the lowest resolution level N is completed (YES in step S49), the initial coefficient encoding process starts, and the NLL subband coefficient stored in the LL subband memory 302 is encoded. Then, it is stored in the NLL subband code memory 304.2 of the output code memory 304 (step S401).
- the layer n to be output is set to the lowest resolution level N (step S402), and the sub-band code memories 304.1 and 304.2 of the output code memory 304 and the NLL sub-band code memory 304.2 are set.
- the pixel extraction process can extract a 2m ⁇ 2 partial area.
- m 2 and 4
- the low-frequency components of each resolution are stored in the nLL sub-band memory.
- the present invention does not limit these storage methods.
- a method may be adopted in which these storage units are shared to reduce the amount of memory used.
- the encoding device and the encoding control according to the present embodiment, after the wavelet operation is performed on the 2m ⁇ 2 pixel region extracted by the pixel extraction process, the LH, HL, and HH coefficients are stored in a small scale such as a register. ⁇ Encoded immediately on high-speed memory (counting memory 303). This enables high-speed encoding and eliminates the need for a real memory for storing the LH, HL, and HH subbands.
- FIG. 8 is a diagram schematically showing encoding and decoding operations according to the first and second embodiments of the present invention.
- a video camera is used as an image input device
- a computer is used as a data processing device 101, 201, 301, and a data storage device 102-104, 220, 203, and 302-304.
- a wired or wireless network communication control device as code output devices 105 and 305.
- the personal computer has a CPU that realizes the functions of image input, wavelet operation, initial coefficient encoding, coefficient extraction, coefficient encoding, and code output.
- image data (0LL) is stored in a semiconductor memory by an image input process.
- the stored image 10 has Hl, H2, ... as pixels, and each pixel is composed of numerical data of Y, Cb, Cr (H1 # Y, Hl # Cb, Hl # Cr, ⁇ ing.
- the input image 10 is subjected to a two-dimensional Haar wavelet transform by a wavelet operation, and the result is stored as nLL, nLH, nHL and nHH subband coefficients Z101-Z104 in the semiconductor memory. Specifically, first, from the origin at the upper left corner,
- the code Z105 thus obtained is stored in the ILL code memory in FIG.
- LH, HL, and HH subbands are encoded in order from the resolution level N.
- the above processing is performed up to the lower right end of the subband, and the obtained code Z106 is stored in the (LH, HL, HH) code memory.
- the personal computer has a CPU that realizes the functions of code input, initial coefficient decoding, coefficient decoding, inverse wavelet operation, and image output.
- code data Z105 and Z106 are stored in the semiconductor memory by a code input process.
- the ILL code and the (1LH, 1HL, 1HH) code are stored.
- the input code Z105 is decoded by the initial coefficient decoding process, and the result is output to the LL sub-band memory in the semiconductor memory.
- (LH, HL, HH) code data Z106 is sequentially supplied to a coefficient decoding process, and the coefficient value is decoded.
- a two-dimensional Haar inverse wavelet operation is performed by a reverse wavelet operation process from the set of the decoded coefficients and the coefficients of the LL subband memory, and the result is output to the 0LL subband memory.
- an inverse wavelet operation is performed using the coefficient values L1, A1, B1, and C1
- the operation results are H1, H2, H3, and Output to H4.
- an inverse wavelet operation is performed using the coefficient values L2, A2, B2, and C2.
- the personal computer in this specific example has a CPU that realizes the functions of image input, pixel extraction, wavelet operation, coefficient coding, initial coefficient coding, and code output.
- image data is stored in a semiconductor memory by an image input process.
- the input image 10 is extracted every 2 ⁇ 2 pixels by a pixel extraction process, supplied to a wavelet operation process, and subjected to wavelet operation and encoding.
- the pixels H1, H2, H3, and H4 are supplied to the wavelet operation process for encoding, and then the pixels H5, H6, H7, and H8 are supplied to the wavelet operation process for encoding.
- the wavelet calculation process when 2 ⁇ 2 pixels are supplied, the pixels are subjected to a two-dimensional Haar wavelet transform to obtain coefficients L1, A1, B1, and C1.
- the coefficient encoding process sequentially encodes A1, B1, and C1 and stores them in the output code memory as (LH, HL, HH) code Z106.
- the code data stored in the semiconductor memory is sorted in order from the resolution level N, and the data in the LL code memory and the (LH, HL, HH) code memory are sequentially transmitted. Output to the road.
- FIG. 9 is a block diagram showing a schematic configuration of an encoding device according to the third embodiment of the present invention.
- the encoding device inputs image data 10 as a representative example of a two-dimensional signal from an image input device (not shown), and outputs a code sequence by wavelet encoding described later.
- the encoding device mainly includes a data processing device 401, an LL subband memory 402, a coefficient memory 403, an output code memory 404, and a pixel thinning map memory 405, and a generated code sequence is a code output device 406.
- the data processing device 401 is a general program control processor such as a CPU, and reads and executes an encoded program group from a program memory (not shown) to execute an inter-pixel bow I map generation process (inter-pixel Bow I map generation means 401 a), pixel thinning extraction Process (pixel decimation extraction means 401b), wavelet transform (here, two-dimensional Haar-single bullet transform) process (wavelet transform means 401c), coefficient encoding process (coefficient encoding means 401d), initial coefficient encoding (initial coefficient encoding) And a code output process (code output means 40 #) can be generated (the details will be described later).
- inter-pixel Bow I map generation means 401 a inter-pixel Bow I map generation means 401 a
- pixel thinning extraction Process pixel decimation extraction means 401b
- wavelet transform here, two-dimensional Haar-single bullet transform
- coefficient encoding process coefficient encoding means 401d
- initial coefficient encoding initial coefficient encoding
- the LL sub-band memory 402, coefficient memory 403, output code memory 404, and pixel thinning map memory 405 each have a force S that can be configured as a separate memory.
- each space is allocated to one semiconductor memory. It is also possible to configure by.
- a register or a cache memory of the CPU can be used as the coefficient memory 403.
- the LL sub-band memory 402 stores the input image data 10 as 0LL sub-band data, and further obtains the two-dimensional Haar wavelet transform of the partial area (2mx2) sequentially extracted by the pixel thinning extraction process. NLL subband coefficients (1 ⁇ n ⁇ N) are stored sequentially, and finally the NLL subband coefficients are stored.
- the pixel thinning-out extraction process sequentially extracts a partial area of 2mx2 pixels from the (n-1) LL subband coefficients, performs a thinning-out process on the partial area, and generates a result. Then, a two-dimensional Haar wavelet transform is performed on it, and nLL subband coefficients are obtained. Partial areas of 2mx2 pixels are sequentially extracted in the scan line direction.
- the coefficient memory 403 stores a set of coefficients ⁇ LH (i, j) of the same spatial coordinates (i, j) in the subbands nLH, nHL, and nHH of one layer obtained by performing the two-dimensional Haar wavelet transform. , HL (i, j), HH (i, j) ⁇ .
- two or more fixed coefficient sets can be stored.
- one or more coefficient sets adjacent to ⁇ LH (i, j), HL (i, j), HH (i, j) ⁇ may be extracted and stored in the coefficient memory 403.
- a plurality of sets of coefficients can be included in the thread.
- the output code memory 404 has a (nLH, nHL, nHH) subband code memory 404.1 and an NLL subband code memory 404.2 that is an initial output code.
- the sub-band code memory 404.1 is a code (nLH, nHL) obtained by coding a set of the same coordinate coefficients sequentially read out to the coefficient memory 403 by a coefficient coding process.
- NHH Store c.
- the NLL sub-band code memory 404.2 stores the sub-band code NLLc obtained by coding the NLL sub-band coefficient by the initial coefficient coding process of the data processing device 401.
- the pixel thinning map memory 405 is generated by a pixel thinning map generation process described later, and stores the resolution level of each part of the image. Therefore, by searching the pixel thinning map memory 405, it is possible to determine whether or not to perform the pixel thinning process on the extracted partial region (details will be described later).
- FIG. 10 is a flowchart showing the encoding control in this embodiment.
- image data 10 is input from an image input device such as a camera or a scanner, and stored in the LL sub-band memory 402 as a 0LL sub-band (step S51).
- the variable n indicating the resolution level (hierarchy) is set to 1
- a partial area of 2m ⁇ 2 pixels corresponding to the coordinates (i, j) is sequentially extracted from the (n_l) LL subband of the LL subband memory 402 (step S54).
- the pixel thinning extraction process reads the resolution level of the coordinates corresponding to the extracted partial area from the pixel thinning map memory 405, and the extracted part is thinned according to the value of the resolution level. It is determined whether the pixel is the target pixel (step S55). If the extracted portion is a thinning target (YES in step S55), pixels are thinned by a predetermined thinning operation (step S56), and a wavelet transform is performed on the thinned pixel values (step S57).
- a wavelet transform is executed using the pixel values of the extracted portion as they are (step S57).
- the thinning operation is not particularly limited. One pixel value in the extraction area may be used as a representative value, or an average of all the extracted pixel values may be taken.
- the extracted 2 ⁇ 2 pixels are pixels to be decimated
- the extracted 2 ⁇ 2 pixels are supplied to the wavelet calculation process as 2 ⁇ 2 pixels having an average value of those pixel values. If the pixel is not the pixel to be decimated, the pixel value of each 2 ⁇ 2 pixel is supplied to the wavelet calculation process as it is.
- the wavelet transform process performs a two-dimensional Haar wavelet transform on the extracted 2m x 2 pixel subregion (step S57), and obtains the resulting LH coefficient nA (i, j) and HL coefficient.
- nB (i, j) and HH coefficient nC (i, j) are stored in coefficient memory 403, respectively, and LL coefficient nL (i, j) is stored in LL subband memory 402 as nLL subband (step S58). .
- the wavelet transform process calls the coefficient code conversion process, reads out the coefficients nA (i, j), nB (i, j), and nC (i, j) from the coefficient memory 403 and Then, the data is stored in the subband code memory 404.1 of the output code memory 404 (step S59).
- the pixel thinning-out extraction process determines whether or not the encoding has been completed for all the coordinates of the layer n (step S60).
- step S502 When the encoding of the subband of the lowest resolution level N is completed (YES in step S502), the initial coefficient encoding process starts, and encodes the NLL subband coefficients stored in the LL subband memory 402. Is stored in the NLL sub-band code memory 404.2 of the output code memory 404 (step S504).
- Steps S504 to S508 are the same as Steps S401 to S405 in FIG. 7, and thus description thereof will be omitted.
- the method of determining the resolution level in the pixel thinning map generation process is not limited.
- the resolution level may be different between a photograph's "natural image” and a character's graph depending on the intention of the image viewer.
- the resolution may be different between an area where the displacement of the adjacent element value in the image is relatively large and an area where the displacement of the adjacent element value is relatively small.
- the resolution of a relatively small area may be set lower than the resolution of a relatively large area.
- the magnitude of the displacement of the adjacent element value can be determined depending on whether the displacement of the adjacent element value is larger or smaller than a preset threshold.
- the encoding processing can be speeded up and the required memory can be reduced as in the first embodiment.
- images can be encoded at different resolution levels for each partial area.
- the processing load caused by the thinning can be kept very small.
- Steps S52 and S55-S56 which are the thinning-related processes described above, can also be applied to the encoding control flow of the first embodiment shown in FIG.
- a pixel thinning map generation step S52 is inserted after the image data input step S11. Further, in the same coordinate coefficient extraction step S15, the same coordinate coefficient extraction process refers to the pixel thinning map, extracts the pixel as it is if there is no need to thin out the pixel, and if it is necessary to thin out the pixel, increases the coefficient to the resolution set for the coordinate. And extract it.
- FIG. 11 is a diagram schematically showing encoding and decoding operations according to the third embodiment of the present invention.
- a video camera is used as an image input device
- a personal computer is used as a data processing device 401
- a semiconductor memory (included in a personal computer) is used as a data storage device 402-405
- a wired or wireless network is used as a code output device 406.
- Communication control device The basic configuration of the present example is the same as that of the second example. The difference lies in that a personal computer as the force data processing device 401 functions to generate a pixel thinning map and extract a pixel thinning.
- image data is stored in a semiconductor memory by an image input process.
- the pixel thinning map generation process analyzes the features such as the edge strength of the input image, separates the photo area from the text area, and sets the pixel resolution for the photo area to resolution level 2 and the text area to resolution level 1. Generate the map Z202.
- the pixel thinning extraction process uses the input image 10 and the pixel thinning map Z202.
- X Extract a pixel area for every two pixels. If the pixel area is a photographic area, thin out the pixels and perform wavelet calculation and encoding.If the pixel area is a character area, perform wavelet calculation and encoding as it is. Do.
- the code string output in this manner is decoded by the decoding device, and a decoded image Z203 in which only the resolution of the photographic region is low is obtained.
- FIG. 12 is a block diagram showing a schematic configuration of a code apparatus according to a fourth embodiment of the present invention.
- the coding device inputs image data 10 as a representative example of a two-dimensional signal from an image input device (not shown) such as a video camera, and outputs a code sequence by wavelet coding described later.
- the encoding device mainly includes a data processing device 501, an LL subband memory 502, a coefficient memory 503, an output code memory 504, a pixel thinning map memory 505, and a subsequent frame data memory 506. Is output to the code output device 507.
- the data processing device 501 is a general program control processor such as a CPU, and reads and executes an encoded program group from a program memory (not shown) to execute a moving area detection process (moving area detecting means 501a). ), Pixel thinning map generating process (pixel thinning map generating means 501b), pixel thinning extracting process (pixel thinning extracting means 501c), ⁇ wavelet transform (here, two-dimensional Haar wavelet transform) process (wavelet transforming means 501d) ), A coefficient encoding process (coefficient encoding means 501e), an initial coefficient encoding process (initial coefficient encoding means 501f), and a code output process (code output means 50lg) can be generated (to be described in detail later). .).
- the LL sub-band memory 502, coefficient memory 503, output code memory 504, pixel thinning map memory 505, and subsequent frame data memory 506 can each be configured as a separate memory.
- one semiconductor memory It is also possible to configure by allocating each space to each.
- a register or a cache memory of the CPU can be used as the coefficient memory 503.
- the LL sub-band memory 502 stores the input image data 10 as 0LL sub-band data Then, the nLL subband coefficients (1 ⁇ n ⁇ N) obtained by performing the two-dimensional Haar wavelet transform on the partial area (2mx 2) sequentially extracted by the pixel extraction process are sequentially stored. NLL subband coefficients are stored.
- the pixel thinning-out extraction process sequentially extracts a partial area of 2 mx 2 pixels from the (n-1) LL subband coefficients, and performs thinning-out processing on the extracted region. Is subjected to a two-dimensional Hall wavelet transform to obtain nLL subband coefficients.
- the partial area (2m X 2) is sequentially extracted in the scan line direction.
- the coefficient memory 503 stores a set of coefficients ⁇ LH (i, j) of the same spatial coordinates (i, j) in the subbands nLH, nHL, and nHH of one layer obtained by performing the two-dimensional Haar wavelet transform. , HL (i, j), HH (i, j) ⁇ .
- two or more fixed coefficient sets can be stored.
- one or more coefficient sets adjacent to ⁇ LH (i, j), HL (i, j), HH (i, j) ⁇ may be extracted and stored in the coefficient memory 503.
- a plurality of sets of coefficients can be stored in the thread.
- the output code memory 504 includes (nLH, nHL, nHH) sub-band code memories 504-1 and NLL sub-band code memories 504.2, which are initial output codes.
- the sub-band code memory 504.1 is a code (nLH, nHL) obtained by coding a set of the same coordinate coefficients sequentially read out to the coefficient memory 503 by a coefficient coding process.
- NHH NLL sub-band code memory 504.2 stores the sub-band code NLLc obtained by coding the NLL sub-band coefficient by the initial coefficient coding process of the data processing device 501.
- the pixel thinning map memory 505 is generated by a pixel thinning map generation process described later, and stores the resolution level of each part of the image. Therefore, by searching the pixel thinning map memory 505, it is possible to determine whether or not to perform the pixel thinning process on the extracted partial region (details will be described later).
- the subsequent frame data memory 506 stores input frame data following the current frame stored in the LL subband memory 502.
- the moving area detection process on the data processing device 501 detects an area (moving area) where a pixel value changes in a subsequent frame from the current input image and the contents of the subsequent frame data memory 506.
- the pixel thinning map generation process generates a coefficient thinning map having a different resolution between the moving area and the other areas based on the information on the detected moving area. Specifically, the moving area is set to low resolution, and the other areas are set to high resolution.
- FIG. 13 is a flowchart showing the encoding control in the present embodiment.
- Step S6 Do when moving image data 10 is input from a moving image input device such as a video camera, frame data to be encoded is stored in the LL subband memory 502 as 0LL subbands, and the subsequent frame data is stored in the subsequent frame data. Stored in the memory 506 (Step S6 Do)
- the moving area detection process detects a moving area based on the 0LL subband data stored in the LL subband memory 502 and the subsequent frame data stored in the subsequent frame data memory 506 (step S62).
- a method of detecting a moving area for example, there is a method of calculating a difference between both frame images.
- a pixel thinning map in which the detected moving area is set to low resolution and the other areas are set to high resolution is generated and stored in the pixel thinning map memory 505 (step S63).
- n indicating a resolution level (hierarchy)
- a partial area of 2m ⁇ 2 pixels corresponding to the coordinates (i, j) is sequentially extracted from the (n_l) LL subband of the LL subband memory 502 (step S54).
- the upper left 2 ⁇ 2 partial area (Hl, H2, H3, H4) is extracted.
- the pixel thinning extraction process performs processing on the extracted partial regions.
- the resolution level of the corresponding coordinates is read from the pixel thinning map memory 505, and it is determined whether or not the extracted portion is a thinning target pixel based on the value of the resolution level (step S66). If the extracted portion is a thinning target (YES in step S66), pixels are thinned by a predetermined thinning operation (step S67), and a wavelet transform is performed on the thinned pixel values (step S68). ). If the extracted portion is not a thinning target (N ⁇ in step S66), the wavelet transform is performed using the pixel values of the extracted portion as they are (step S68).
- the thinning operation is not particularly limited. One pixel value in the extraction area may be used as a representative value, or an average of all the extracted pixel values may be taken.
- the extracted 2 ⁇ 2 pixels are pixels to be decimated
- the extracted 2 ⁇ 2 pixels are supplied to the wavelet calculation process as 2 ⁇ 2 pixels having an average value of those pixel values. If the pixel is not the pixel to be decimated, the pixel value of each 2 ⁇ 2 pixel is supplied to the wavelet calculation process as it is.
- the wavelet transform process performs a two-dimensional Haar wavelet transform on the extracted 2m x 2 pixel partial area (step S68), and obtains the resulting LH coefficient nA (i, j) and HL coefficient.
- the nB (i, j) and the HH coefficient nC (i, j) are stored in the coefficient memory 503, respectively, and the LL coefficient nL (i, j) is stored in the LL subband memory 502 as an nLL subband (step S69). .
- the wavelet transform process calls the coefficient encoding process, reads out the coefficient sets nA (i, j), nB (i, j), nC (i, j) from the coefficient memory 503 and sets the variable length.
- the encoding is performed, and stored in the sub-band code memory 504.1 of the output code memory 504 (step S70).
- the pixel thinning-out extraction process determines whether or not encoding has been completed for all coordinates of the layer n. A determination is made (step S71).
- step S71 If all the coding in the layer n has not been completed (N ⁇ in step S71), the extraction coordinates (i, j) are shifted by a predetermined number in the scan line direction (step S72). Steps S65 and S72 are repeated until all encoding in n is completed.
- step S73 When the encoding of the subband of the lowest resolution level N is completed (YES in step S73), the initial coefficient encoding process starts, and the N stored in the LL subband memory 502 is set.
- the LL subband coefficient is encoded, and the NLL subband code memory 504 of the output code memory 504 is encoded.
- steps S75-S79 are the same as steps S401-S405 in Fig. 7, and a description thereof will be omitted.
- the subsequent frame information used in the present embodiment may be image data of a subsequent frame, or may be coordinate information of an update area in which a signal value is updated.
- image data is used as the succeeding frame information
- the moving area can be detected by taking the difference between the current frame image and the succeeding frame image.
- the coordinate information of the update area is used, the update area can be used as it is as the moving area.
- a plurality of frames may be used instead of a single frame.
- a period during which the value of the pixel changes in each partial region is obtained, and the resolution of the pixel thinning map is set based on the change period. More specifically, the resolution is lower in a region where the change period of the pixel is long, that is, in a region where the pixel changes over a long time.
- steps S61 S63 and S66 S67 which are the moving area detection and thinning-related processing in the fourth embodiment described above, can be applied to the encoding control flow of the first embodiment shown in FIG. .
- the coding apparatus is similar to the first embodiment.
- a region overwritten by a subsequent frame image is detected as a moving region, so that the moving region has a low resolution. It is possible to generate a pixel thinning map in which the other regions are set to a high resolution, thereby reducing the amount of code in the moving region and improving the frame rate.
- the resolution of the moving area will decrease, but since the moving area is an area that is overwritten in the subsequent frame, even if it is displayed temporarily at a low resolution, viewers will not be able to see It is difficult to recognize quality deterioration.
- FIG. 14 is a diagram schematically showing a console screen for describing a moving area detection and a thinning-out operation according to the fourth embodiment of the present invention.
- a video camera is used as an image input device
- a workstation CPU is used as a data processing device 501
- a semiconductor storage device (included in a workstation) is used as a data storage device 502-506, and
- a code output device is used as a code output device. It has a communication control device connected to the network.
- the workstation CPU can implement the functions of moving area detection, pixel thinning map generation, pixel thinning extraction, ⁇ Ablet transform, coefficient coding, initial coefficient coding, and code output. .
- a console screen of a workstation is assumed as a display means for schematically showing the moving area detection and thinning processing of an image.
- Fig. 14 an update area of a frame screen is sequentially supplied, and the The state of transmission after transmission is schematically shown.
- Z300 indicates the state of the console screen in frame f
- Z3000 indicates the update area on screen Z300 in frame f
- Z301 indicates the state of the console screen in frame (f + 1)
- Z3010 indicates the update area on screen Z301 in frame (f + 1).
- Z302 indicates the state of the console screen at frame (f + 2)
- Z3020 indicates the update area.
- the workstation has previously acquired the coordinates of the update area of the succeeding frame (f + 1). That is, Z3001 indicates the position of the update area of the subsequent frame (f + 1) on the screen Z300 of the frame f, and similarly, Z3011 indicates the subsequent frame (f + 1) of the frame (f + 1) on the screen Z301. +2) indicates the coordinates of the update area. Also, In the frame following frame (f + 2), it is assumed that screen Z302 has not been updated.
- the operation in frame f is as follows. First, the workstation acquires the input image Z3000 and the subsequent frame information Z3001. The overlapping area of Z3000 and Z3001 is detected as a moving area by the moving area detection process, and the pixel thinning map Z310 that sets the moving area to resolution level 2 and the other areas to resolution level 1 by the pixel thinning map generation process. Generated.
- a pixel is extracted every 2 X 2 pixels based on the input image Z3000 and the pixel thinning map Z310, supplied to the wavelet operation process, and performs wavelet transform and encoding. Pixels are decimated and encoded for the moving area of resolution level 2. By decoding the code obtained in this way and performing drawing in the update area, a display image Z320 in which only the resolution of the moving area is reduced is obtained. The above is the operation in frame f.
- FIG. 15 is a block diagram illustrating a schematic configuration of an encoding device according to a fifth embodiment of the present invention.
- the basic configuration of the present embodiment is the same as that of the second embodiment, except that the data processing device 9001 can generate a coefficient quantization map generation process and a coefficient quantization process. It differs in that it has.
- the data processing device 901 includes a coefficient quantization map generation process (coefficient quantization map generation means 901a), a pixel extraction process (pixel extraction (element extraction) means 901b), and an ⁇ wavelet transform (here, 2D Haar wavelet transform) process (wavelet transformer 901c), coefficient quantization process (coefficient quantizer 901d), coefficient encoding process (Coefficient encoding means 901e), an initial coefficient encoding process (initial coefficient encoding means 901f), and a code output process (code output means 901g).
- coefficient quantization map generation means 901a coefficient quantization map generation means 901a
- a pixel extraction process pixel extraction (element extraction) means 901b
- an ⁇ wavelet transform here, 2D Haar wavelet transform
- coefficient quantizer 901d coefficient quantization process
- coefficient encoding means 901e coefficient encoding process
- an initial coefficient encoding process initial coefficient encoding means 901f
- code output means 901g code output means
- the coefficient quantization map memory 905 stores a coefficient quantization map, which is a quantization parameter for each part of the image. Therefore, by searching the coefficient quantization map memory 905, it is possible to obtain the accuracy of the coefficient quantization for the partial region extracted in the pixel extraction process (details will be described later).
- the coefficient quantization map is generated by a coefficient quantization map generation process.
- FIG. 16 is a flowchart showing the encoding control in the present embodiment.
- image data 10 is input from an image input device such as a camera or a scanner, and stored in the LL sub-band memory 902 as 0LL sub-band (step S801).
- the coefficient quantization map generation process on the data processing device 901 reads the 0LL sub-band data when an image is input, and determines the coefficient corresponding to each spatial coordinate position and each resolution level by the quantization accuracy.
- a coefficient quantization map indicating whether or not to perform encoding by thinning out is generated and stored in the coefficient quantization map memory 905 (step S802).
- n indicating the resolution level (hierarchy)
- a partial area of 2m ⁇ 2 pixels corresponding to the coordinates (i, j) is sequentially extracted from the (n ⁇ 1) LL subband of the LL subband memory 902 (step S804) ).
- the wavelet transform process performs a two-dimensional Haar wavelet transform on the extracted 2m x 2 pixel partial area (step S805), and obtains the resulting LH coefficient nA (i, j) and HL coefficient.
- the nB (i, j) and the HH coefficient nC (i, j) are stored in the coefficient memory 903, respectively, and the LL coefficient nL (i, j) is stored in the LL subband memory 902 as an nLL subband (step S806).
- the coefficient quantization process uses the LH coefficient nA (i, j), HL coefficient nB (i, j) and And the HH coefficient nC (i, j) are read. Further, from the coefficient quantization map memory 905, a coefficient quantization parameter corresponding to the space coordinate and the resolution level is read, and the LH coefficient nA (i, j) and the HL coefficient nB (i , j) and the HH coefficient nC (i, j) are quantized. The contents of the coefficient memory 903 are updated by each quantized coefficient (step S807).
- steps S808-S817 are the same as steps S46-S50, S401, and S405 in Fig. 7, and thus description thereof will be omitted.
- reference numeral 904 denotes an output code memory
- 904.1 denotes an (nLH, nHL, nHH) subband code memory
- 904.2 denotes an NLL subband code memory
- 906 denotes a code output device.
- the coefficient quantization map may be generated so that the quantization accuracy of the coefficient differs between the photograph area and the character area. For example, if a single image contains both text areas with gray levels and photographs with varying colors and gray levels, detecting the displacement of the P-tangent element value, By analyzing the features, the photograph area and the character area can be distinguished. Then, by generating the coefficient quantization map so that the quantization accuracy of the photograph region is lower than that of the character region, it is possible to reduce the code amount while suppressing visual quality deterioration.
- the quantization accuracy of the coefficient may be different between an area where the displacement of the adjacent element value in the image is relatively large and an area where the displacement of the adjacent element value is relatively small.
- the quantization precision of a relatively small area may be set lower than that of a relatively large area.
- the magnitude of the displacement of the neighboring element value can be determined depending on whether the displacement of the neighboring element value is larger or smaller than a preset threshold.
- the coding processing can be speeded up and the required memory can be reduced as in the first embodiment.
- images can be encoded with different quantization accuracy for each subregion.
- the waiting time can be reduced.
- quality control can be performed in finer units than in the third embodiment.
- the resolution when reducing the quality of the non-interest area, the resolution is reduced by a power of 2 such as 2 or 4. For this reason, it was necessary to reduce the resolution by half even when it was desired to slightly lower the quality.
- the quantization accuracy by controlling the quantization accuracy, it is possible to finely control the quantization distortion such as mosquito noise.
- FIG. 17 is a diagram schematically showing an encoding operation according to the fifth embodiment of the present invention.
- a video camera is used as an image input device
- a personal computer is used as a data processing device 901
- a semiconductor memory (included in a personal computer) is used as a data storage device 902-905
- a wired or wireless device is used as a code output device 906. It is assumed that a network communication control device is provided.
- the basic configuration of this example is the same as that of the second example, except that a personal computer as the data processing device 901 has a function of generating a coefficient quantization map and a coefficient quantization.
- the coefficient quantization process based on the information of the quantization step obtained from the coefficient quantization map Z402, if the pixel area is a photograph area, the coefficient value Z403.1 is quantized in six steps, and the coefficient value Z403. Get two. If the pixel region is a character region, the pixel is quantized in three steps to obtain a coefficient value Z403.3.
- the quantized coefficients are encoded by a coefficient encoding process. [0200]
- the code string output in this way is decoded by the decoding device, and an image in which only the coefficients of the photographic region are roughly quantized is obtained.
- FIG. 18 is a block diagram showing a schematic configuration of a code encoder according to a sixth embodiment of the present invention.
- the basic configuration of this embodiment is the same as that of the fifth embodiment, except that the data processing device 1001 can generate a moving area detection process, and that it has a subsequent frame data memory 1006.
- the data processing device 1001 performs a moving region detection process (moving region detection means 1001a), a coefficient quantization map generation process (coefficient quantization map generation means 1001b), a pixel extraction process (pixel extraction (element extraction ) Means 1001c), wavelet transform (here, 2D Haar wavelet transform) process (wavelet transform means lOOld), coefficient quantization process (coefficient quantizer lOOle), coefficient encoding process (coefficient encoding means 1001f) , An initial coefficient encoding process (initial coefficient encoding means 1001g), and a code output process (code output means 1001h).
- a moving region detection process moving region detection means 1001a
- coefficient quantization map generation means 1001b coefficient quantization map generation means 1001b
- a pixel extraction process pixel extraction (element extraction ) Means 1001c
- wavelet transform here, 2D Haar wavelet transform
- coefficient quantizer lOOle coefficient quantization process
- coefficient encoding means 1001f coefficient encoding
- the moving area detection process on the data processing device 1001 is based on the current input image and the contents of the subsequent frame data memory 1006 where the pixel value changes in the subsequent frame (moving area). ) Is detected.
- the coefficient quantization map generation process generates a coefficient quantization map having a different quantization precision between the moving region and the other region based on the detected information on the moving region. Specifically, the moving region is set to a coarse quantization accuracy, and the other regions are set to a fine quantization accuracy.
- FIG. 19 is a flowchart showing the encoding control in this embodiment.
- step S901 when moving image data 10 is input from a moving image input device such as a video camera, frame data to be encoded is stored in the LL subband memory 1002 as a 0LL subband, and the subsequent frame data is stored in a subsequent frame.
- the data is stored in the data memory 1006 (step S901).
- the moving area detection process is based on the 0LL subband data stored in the LL subband memory 1002 and the subsequent frame data stored in the subsequent frame data memory 1006.
- a moving area is detected (step S902).
- a method of detecting a moving area for example, there is a method of calculating a difference between both frame images.
- the coefficient quantization map generation process generates a coefficient quantization map in which the detected moving region is set to coarse quantization accuracy and the other regions are set to fine power, fineness, and quantization accuracy. It is stored in the child map memory 1005 (step S903).
- steps S904 to S918 are the same as steps S803 and S817 in Fig. 16, and description thereof will be omitted.
- 1003 is a coefficient memory
- 1004 is an output code memory
- 1004.1 is an (nLH, nHL, nHH) subband code memory
- 1004.2 is an NLL subband code memory
- 1007 is a code output device. is there.
- the subsequent frame information used in the present embodiment may be image data of a subsequent frame, or may be coordinate information of an update area in which a signal value is updated.
- image data is used as the succeeding frame information
- the moving area can be detected by taking the difference between the current frame image and the succeeding frame image.
- the coordinate information of the update area is used, the update area can be used as it is as the moving area.
- a plurality of frames may be used instead of a single frame.
- a period during which the value of the pixel changes in each partial area is obtained, and the quantization accuracy of the coefficient quantization map is set based on the change period. More specifically, the quantization accuracy is set lower in a region where the pixel change period is long, that is, in a region where the pixel changes over a long time.
- the encoding process can be speeded up and the required memory can be reduced as in the first embodiment. It is possible to reduce the amount of code in the moving region and improve the frame rate.
- quality control can be performed in finer units than in the fourth embodiment.
- the resolution is reduced by a power of 2, such as 1/2 or 1/4. For this reason, it was necessary to reduce the resolution by half even when it was desired to slightly lower the quality.
- the sixth embodiment by controlling the quantization accuracy, it is possible to finely control the quantization distortion such as mosquito noise.
- the basic configuration of this example is the same as that of the fourth example, but differs in that it has a function of a workstation S as the data processing device 1001, a function of generating a coefficient quantization map, and a function of coefficient quantization.
- the quality of the moving area can be reduced in the same manner as in the fourth specific example, but the image is displayed with a lower quantization precision than in the fourth specific example. You.
- FIG. 20 is a block diagram showing a schematic configuration of an encoding device according to the seventh embodiment of the present invention.
- the basic configuration of the present embodiment is the same as that of the first embodiment, except that the data processing device 1101 can generate the moving region detection process, the coefficient quantization map generation process, the coefficient quantization process, and the coefficient quantization process. It is different in that it has a generalized map memory 1105 and a subsequent frame data memory 1106.
- the data processing device 1101 includes a moving region detection process (moving region detection means 1101a), a coefficient quantization map generation process (coefficient quantization map generation means 1101b), a wavelet transform (here, a two-dimensional Haar Wavelet transform) process (wavelet transform means 1101c), same coordinate coefficient extraction process (same coordinate coefficient extraction means l lOld), coefficient quantization process (coefficient quantization means 1101e), initial coefficient encoding process (initial coefficient encoding means) 110 ⁇ ), a coefficient encoding process (coefficient encoding means l lOlg), and a code output process (code output means l lOlh) can be generated.
- the moving area detection process on the data processing device 1101 is similar to that of the fourth embodiment, in the area where the pixel value changes in the subsequent frame (moving area) from the current input image and the contents of the subsequent frame data memory 1106. ) Is detected.
- the coefficient quantization map generation process generates a coefficient quantization map having a different quantization precision between the moving region and the other region based on the information of the detected moving region. Specifically, the moving region is set to a coarse quantization accuracy, and the other regions are set to a fine quantization accuracy.
- the generated coefficient quantization map is stored in the coefficient quantization map memory 1105.
- the LH coefficient nA (i, j), the HL coefficient nB (i, j), and the HH coefficient nC (i, j) are read from the coefficient memory 1003. Further, from the coefficient quantization map memory 1105, a coefficient quantization parameter corresponding to the spatial coordinate 'resolution level is read, and the LH coefficient nA (i, j) and the HL coefficient nB up to the quantization precision indicated by the parameter. Quantize (i, j) and HH coefficient nC (i, j).
- FIG. 21 is a flowchart showing the encoding control in the present embodiment.
- the frame data to be encoded is stored in the subband memory 1102 as a 0LL subband, and the subsequent frame data is stored in the subsequent frame data memory. It is stored in 1106 (step S1 001).
- the wavelet transform process on the data processing device 101 executes N-stage two-dimensional Haar wavelet transform on the 0LL subband data read from the subband memory 1102, and performs the transform.
- the initial coefficient coding process reads the initial coefficient NLL from the subband memory of the layer N, and performs variable-length coding using a run-length Huffman code or an LZW code.
- the codes NLLc (0, 0), NLLc (l, 0), ..., NLLc (p, q) are stored in the NLL subband code memory 1104.1 of the output code memory 104.
- the code 1J stored in the output code memory 1104 is read and output to the code output device 1107 (step S1004).
- the next coding target is the LH, HL, and HH subbands.
- the variable n indicating the resolution level (hierarchy) is set to N
- the moving region detection process and the coefficient quantization map generation process are activated.
- the moving area detection process detects a moving area based on the 0LL subband data stored in the LL subband memory 1102 and the subsequent frame data currently stored in the subsequent frame data memory 1106 (step S1006).
- a method of detecting a moving area for example, there is a method of calculating a difference between two frame images.
- the coefficient quantization map generation process generates a coefficient quantization map in which the detected moving region is set to coarse quantization accuracy and the other regions are set to fine quantization accuracy, and the coefficient quantization map memory 1105 (Step S1007).
- the same coordinate coefficient extraction process performs the same set of coefficients (nA (i, j), nB) of the same spatial coordinates (i, j) in the subband nLH, nHL, nHH of the hierarchy n from the subband memory 1102. (i, j), nC (i, j) ⁇ is read and stored in the coefficient memory 1103 (step S1008).
- a coefficient quantization process is activated to perform quantization.
- the coefficient set nA (i, j), nB (i, j), nC (i, j) is read from the coefficient memory 1103, and further, the corresponding space is read from the coefficient quantization map memory 1105.
- the coefficient quantization parameter corresponding to the coordinate 'resolution level is read, and the LH coefficient nA (i, j), HL coefficient nB (i, j), and HH coefficient nC (i, j) is quantized.
- the contents of the coefficient memory 1103 are updated by each quantized coefficient (step S1009).
- the coefficient encoding process reads the coefficient sets nA (i, j), nB (i, j), nC (i, j) from the coefficient memory 1103, performs variable length coding, and outputs the code. It is stored in the subband code memory 1104.2 of the memory 1104 (step S1010).
- the same coordinate coefficient extraction process determines whether or not encoding has been completed for all the coordinates of the layer n (step S 1011).
- step S1011 Unless all encoding in the layer n has been completed (NO in step S1011) Then, the extraction coordinates (i, j) are shifted by a predetermined number in the scan line direction (step S1012), and the above steps S1008 to S1012 are repeated until all the encoding in the layer n is completed.
- step S1011 When all encoding on the layer n is completed (YES in step S1011), the code output process reads the code string stored in the output code memory 1104 and outputs it to the code output device 1107 (step S1013). ).
- the moving region detection process and the coefficient quantization map generation process are activated before the LH, HL, and HH subband encoding of each resolution level. These processes may be started at a certain resolution level to suppress noise, or these processes may be started at a certain resolution level or higher.
- the encoding process can be speeded up and the required memory can be reduced as in the first embodiment. It is possible to reduce the amount of code in the moving region and improve the frame rate.
- the code amount can be further reduced and the frame rate can be further improved.
- the reasons are the following two points.
- the LH, HL, and HH subbands encoded at each resolution are output as is to a code output device 1107 such as a communication channel.
- the code output device 1107 is slower than the processing speed of the data processing device 1101, so that a lot of waiting time is required for outputting the code.
- the screen is updated during the code output period, it can be detected as a moving area immediately before the next resolution level is encoded.
- a moving region can be newly detected from a subsequent frame, and the code amount can be further reduced.
- the code amount control method is a method based on quantization.
- the same frame rate improvement effect can be achieved by using the code amount control by resolution truncation as in the fourth embodiment. .
- FIG. 22 is a diagram schematically showing a console screen for explaining the motion region detection and the coefficient quantization operation according to the seventh embodiment of the present invention.
- the basic configuration in this specific example is the same as that of the fourth specific example, except that a workstation force coefficient quantization map generation process and a coefficient quantization process, which are the data processing device 1101, can be generated.
- Fig. 22 schematically illustrates a state in which the update areas of the frame screen are sequentially supplied, encoded, and transmitted.
- Z501, Z502, and Z503 show the console screen of the workstation in frames f, (f + 1), and (f + 2), respectively.
- Z5010, Z5020, Z5030 indicate the area to be encoded by frames f, (f + 1), (f + 2), and Z5031 is newly supplied in frame (f + 2). The following frame area is shown.
- the workstation obtains input image Z501, performs wavelet transform on region Z5010 of input image Z501, and obtains each frequency subband Z510. Furthermore, Z511, which is the lowest resolution LL component, is variable-length coded and output to the communication channel.
- FIG. 23 is a block diagram of a bidirectional communication terminal incorporating a coding and decoding device according to the eighth embodiment of the present invention.
- the terminal is provided with a CPU 601 as a program control processor, and is connected to the cache memory 602 and the main memory 603 via an internal bus.
- the internal bus is further connected to an external bus 604 via a port.
- the external bus 604 has a memory 605 storing necessary programs, a data memory 606, an interface 607 for connecting a camera 608, and an interface 609 for connecting a display 610. , And an interface 611 connecting the communication control unit 612 and the like.
- the communication control unit 612 is connected to the network 613. If it is a mobile phone, the communication control unit 612
- the network 613 is a mobile communication network including a communication unit and a channel control unit.
- the program memory 605 stores an encoding program, a decoding program, a main program for controlling the overall operation of the communication terminal, and the like.
- the encoding program is represented by the flowchart shown in FIG. 3, FIG. 7, FIG. 10, FIG. 13, FIG. 16, FIG. 19, or FIG. 21 described in the first to seventh embodiments. Is represented by the flowchart shown in FIG.
- the above-described processes of the encoding program and the decoding program are executed under the control of the process by the main program.
- the data memory 606 includes a sub-band memory 606.1 and an input / output encoded data memory 606.
- the sub-band memory 606.1 is, for example, the sub-band memory 102 in FIG. 1 or the LL sub-band memory 302 in FIG.
- Input / output coded data memory 606.2 includes, for example, an output code memory such as output code memory 104/304 in FIGS. 1 and 6, and an input code memory such as the 0LL subband memory in FIG.
- the camera 608 corresponds to the image input device in the encoding device
- the display 610 corresponds to the image output device 204 in the decoding device.
- the image data captured by the camera is subjected to wavelet transform and encoding as described above, and the code sequence is transmitted to the other terminal via the network.
- the code string of the image data received from the other party is decoded and subjected to inverse wavelet transform as described above and displayed on the display 610.
- the detailed operation is as described above, and in any case, the cache memory 602 can be used as a coefficient memory.
- the encoding device and the decoding device according to the present invention can be realized by executing each control program on the CPU as described above, but can also be realized by hardware.
- FIG. 24 is a block diagram showing an example of the encoding device according to the ninth embodiment of the present invention.
- Image data input by an image input device 701 such as a camera is stored in an image memory 702, and is sequentially subjected to wavelet transform by the wavelet transform unit 703 as described above.
- the obtained subband coefficients LL, LH, HL, HH are stored in the subband memory 704, and the initial coefficients NLL are coded by the initial coefficient coding unit 705 and stored in the output code memory 709.
- the same coordinate coefficient extraction unit 706 extracts a coefficient set of the same coordinates of the other subband coefficients LH, HL, and HH from the subband memory 704, and stores one or more constant coefficient sets in the high-speed register 707. Store.
- the coefficient set stored in the register 707 is encoded by the coefficient encoding unit 708 and stored in the output code memory 709. In this way, the NLL code and the code of the same coordinate coefficient set stored in the output code memory 709 are sequentially read and output by the code output unit 710.
- FIG. 25 is a block diagram showing an example of the decoding device according to the ninth embodiment of the present invention.
- the code string transmitted by the coding device is input by a code input device 801 and stored in an input code memory 802.
- the initial coefficient NLL is decoded by the initial coefficient decoding unit 803 and stored in the subband memory 804.
- the coefficient decoding unit 805 sequentially decodes one or more fixed number of subband coefficient sets nLH, nHL, nHH from the input code, and stores them in the register 806.
- the wavelet inverse transform unit 807 reads the coefficient set from the register 806, performs wavelet inverse transform, and stores the result in the subband memory 804.
- the image output unit 808 thereafter outputs image data.
- the processing target is particularly image data, but it goes without saying that the present invention can be applied to two-dimensional data other than image data. not.
- processing program for realizing the present invention as described above can be stored in a recording medium such as a magnetic disk or a CD-ROM. Both the processing program and the recording medium storing the processing program are included in the scope of the present invention.
- the encoding / decoding device can be applied to applications in which high-definition images are gradually distributed according to the resolution of a display device or the operation / request of a viewer.
Abstract
Description
Claims
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006510166A JP4665898B2 (ja) | 2004-02-23 | 2004-12-28 | 2次元信号の符号化/復号方法および装置 |
EP04807959A EP1722572B1 (en) | 2004-02-23 | 2004-12-28 | 2-dimensional signal encoding/decoding method and device |
US10/590,245 US8031951B2 (en) | 2004-02-23 | 2004-12-28 | 2-dimensional signal encoding/decoding method and device |
CN2004800428269A CN1943243B (zh) | 2004-02-23 | 2004-12-28 | 二维信号编码/解码方法和设备 |
US13/226,116 US8265402B2 (en) | 2004-02-23 | 2011-09-06 | 2 dimensional signal encoding/decoding method and device |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004-045659 | 2004-02-23 | ||
JP2004045659 | 2004-02-23 | ||
JP2004-324875 | 2004-11-09 | ||
JP2004324875 | 2004-11-09 |
Related Child Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/590,245 A-371-Of-International US8031951B2 (en) | 2004-02-23 | 2004-12-28 | 2-dimensional signal encoding/decoding method and device |
US13/226,116 Division US8265402B2 (en) | 2004-02-23 | 2011-09-06 | 2 dimensional signal encoding/decoding method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2005081539A1 true WO2005081539A1 (ja) | 2005-09-01 |
Family
ID=34889354
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2004/019604 WO2005081539A1 (ja) | 2004-02-23 | 2004-12-28 | 2次元信号の符号化/復号方法および装置 |
Country Status (5)
Country | Link |
---|---|
US (2) | US8031951B2 (ja) |
EP (1) | EP1722572B1 (ja) |
JP (1) | JP4665898B2 (ja) |
CN (1) | CN1943243B (ja) |
WO (1) | WO2005081539A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008288841A (ja) * | 2007-05-17 | 2008-11-27 | Sony Corp | 符号化装置および方法 |
WO2011013192A1 (ja) * | 2009-07-27 | 2011-02-03 | 株式会社東芝 | 画像符号化方法および画像復号方法 |
US8218645B2 (en) | 2009-07-17 | 2012-07-10 | Nec Corporation | Method and apparatus for encoding and decoding by wavelet transform |
US8395634B2 (en) | 2007-05-17 | 2013-03-12 | Sony Corporation | Method and apparatus for processing information |
EP2942935A1 (en) * | 2006-07-13 | 2015-11-11 | NEC Corporation | Wavelet-based image encoding and decoding device and encoding method and decoding method |
Families Citing this family (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9521431B2 (en) * | 2007-01-03 | 2016-12-13 | Numeri Ltd. | Method and a system for wavelet based processing |
US8619874B2 (en) * | 2007-04-13 | 2013-12-31 | Apple Inc. | Method and system for video encoding and decoding |
US8582656B2 (en) * | 2007-04-13 | 2013-11-12 | Apple Inc. | Method and system for video encoding and decoding |
CN101543038B (zh) * | 2007-05-17 | 2013-01-02 | 索尼株式会社 | 信息处理装置和方法 |
JP5005482B2 (ja) * | 2007-09-26 | 2012-08-22 | 株式会社日立国際電気 | パターン検索方法及び装置 |
JP5092722B2 (ja) * | 2007-12-07 | 2012-12-05 | ソニー株式会社 | 画像処理装置、画像処理方法およびプログラム |
US8768064B2 (en) * | 2008-09-01 | 2014-07-01 | Nec Corporation | Image feature extraction device |
US8755515B1 (en) * | 2008-09-29 | 2014-06-17 | Wai Wu | Parallel signal processing system and method |
JP5263621B2 (ja) * | 2009-09-24 | 2013-08-14 | ソニー株式会社 | 画像処理装置および方法 |
JP2011180792A (ja) * | 2010-03-01 | 2011-09-15 | Sony Corp | 画像処理装置及び画像処理方法、並びにコンピューター・プログラム |
JP6160615B2 (ja) * | 2012-07-19 | 2017-07-12 | 日本電気株式会社 | ウェーブレット変換符号化/復号方法および装置 |
CN103310411B (zh) * | 2012-09-25 | 2017-04-12 | 中兴通讯股份有限公司 | 一种图像局部增强方法和装置 |
FR2996033B1 (fr) * | 2012-09-26 | 2015-10-30 | Commissariat Energie Atomique | Dispositif de decomposition d'images par transformee en ondelettes |
CN103152573A (zh) * | 2013-03-15 | 2013-06-12 | 惠州Tcl移动通信有限公司 | 一种移动终端与智能电视间图像帧传输的方法及系统 |
JP2014203178A (ja) * | 2013-04-02 | 2014-10-27 | 株式会社東芝 | コンテンツ配信システムおよびコンテンツ配信方法 |
CN103400128B (zh) * | 2013-08-09 | 2016-12-28 | 深圳市捷顺科技实业股份有限公司 | 一种图像处理方法及装置 |
US9049432B1 (en) * | 2014-07-04 | 2015-06-02 | Primacomp, Inc. | Motion differential set partition coding for color image sequence compression |
US10547825B2 (en) * | 2014-09-22 | 2020-01-28 | Samsung Electronics Company, Ltd. | Transmission of three-dimensional video |
US10003822B2 (en) | 2016-02-10 | 2018-06-19 | Primacomp, Inc. | Error-resilient coder of image sequences and video |
EP3291452B1 (en) * | 2016-08-30 | 2020-03-25 | Alcatel Lucent | Encoding and decoding of vectoring coefficients with differential encoding size |
CA3048419C (en) * | 2018-02-27 | 2020-10-20 | Zetane Systems Inc. | Scalable transform processing unit for heterogeneous data |
US11044474B2 (en) * | 2019-02-18 | 2021-06-22 | Glenfly Technology Co., Ltd. | Image compression method and image processing system |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0316490A (ja) * | 1989-06-14 | 1991-01-24 | Matsushita Electric Ind Co Ltd | 動画像符号化装置 |
JPH09148938A (ja) * | 1995-11-21 | 1997-06-06 | Seiko Epson Corp | データ圧縮・伸長方法及びその装置 |
JPH1169164A (ja) * | 1997-08-21 | 1999-03-09 | Toshiba Corp | 画像符号化方法および画像符号化装置および画像復号化装置および画像形成装置 |
JPH11317950A (ja) * | 1998-04-30 | 1999-11-16 | Sony Corp | データ符号化装置、データ符号化方法及びデータ伝送方法 |
JP2000004364A (ja) * | 1998-06-15 | 2000-01-07 | Ricoh Co Ltd | 符号化/復号化装置 |
JP2000316174A (ja) * | 1999-04-30 | 2000-11-14 | Ricoh Co Ltd | カラー画像圧縮方法およびカラー画像圧縮装置 |
JP2001025018A (ja) * | 1999-07-06 | 2001-01-26 | Ricoh Co Ltd | ウェーブレット変換装置、符号化復号化装置、ウェーブレット変換処理方法、及び、記録媒体 |
JP2001352543A (ja) * | 2000-06-07 | 2001-12-21 | Canon Inc | 画像符号化システム及び画像符号化装置及び方法及び記憶媒体 |
JP2002016808A (ja) * | 1999-05-10 | 2002-01-18 | Ricoh Co Ltd | 符号化装置、復号化装置、符号化方法、復号化方法、およびその方法を実行するためのコンピュータが読取可能な記録媒体 |
JP2002232721A (ja) * | 2001-01-31 | 2002-08-16 | Canon Inc | 画像符号化装置及びその方法並びに記憶媒体 |
JP2003204439A (ja) * | 2002-01-08 | 2003-07-18 | Canon Inc | 画像符号化装置及び画像符号化方法 |
JP2003274190A (ja) * | 2002-03-12 | 2003-09-26 | Canon Inc | 静止画像符号化方式 |
JP2003274185A (ja) * | 2002-03-19 | 2003-09-26 | Sanyo Electric Co Ltd | 画像処理方法とその方法を利用可能な画像符号化装置 |
Family Cites Families (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5861921A (en) * | 1993-03-29 | 1999-01-19 | Canon Kabushiki Kaisha | Controlling quantization parameters based on code amount |
JPH11298897A (ja) | 1998-04-14 | 1999-10-29 | Sumitomo Electric Ind Ltd | 画像データ伝送方法および画像データ伝送装置 |
US6763139B1 (en) * | 1998-07-03 | 2004-07-13 | Canon Kabushiki Kaisha | Image coding method and apparatus for localized decoding at multiple resolutions |
US6778709B1 (en) * | 1999-03-12 | 2004-08-17 | Hewlett-Packard Development Company, L.P. | Embedded block coding with optimized truncation |
US6477280B1 (en) * | 1999-03-26 | 2002-11-05 | Microsoft Corporation | Lossless adaptive encoding of finite alphabet data |
JP2000316152A (ja) | 1999-05-06 | 2000-11-14 | Ricoh Co Ltd | 符号化方法、復号方法、符号化装置及び復号装置 |
US6788811B1 (en) * | 1999-05-10 | 2004-09-07 | Ricoh Company, Ltd. | Coding apparatus, decoding apparatus, coding method, decoding method, amd computer-readable recording medium for executing the methods |
US6535644B1 (en) | 1999-07-01 | 2003-03-18 | Koninklijke Philips Electronics N.V. | Hierarchical foveation based on wavelets |
AUPQ156199A0 (en) * | 1999-07-12 | 1999-08-05 | Canon Kabushiki Kaisha | Method and apparatus for re-ordering of a bit stream |
US7065253B2 (en) * | 1999-09-03 | 2006-06-20 | Intel Corporation | Wavelet zerotree coding of ordered bits |
US7421136B2 (en) * | 1999-11-24 | 2008-09-02 | Ge Medical Systems Information Technologies Inc. | Image tessellation for region-specific coefficient access |
US6549674B1 (en) * | 2000-10-12 | 2003-04-15 | Picsurf, Inc. | Image compression based on tiled wavelet-like transform using edge and non-edge filters |
EP1297709A1 (en) * | 2000-06-14 | 2003-04-02 | Koninklijke Philips Electronics N.V. | Color video encoding and decoding method |
US20030067476A1 (en) * | 2001-10-04 | 2003-04-10 | Eastman Kodak Company | Method and system for displaying an image |
JP3922919B2 (ja) * | 2001-12-11 | 2007-05-30 | 株式会社リコー | 静止画像伸長装置及び静止画像伸長方法 |
EP1349393A1 (en) | 2002-03-15 | 2003-10-01 | Ricoh Company | Image compression device, image decompression device, image compression/decompression device, program for executing on a computer to perform functions of such devices, and recording medium storing such a program |
JP2004254298A (ja) * | 2003-01-30 | 2004-09-09 | Ricoh Co Ltd | 画像処理装置、プログラム及び記憶媒体 |
-
2004
- 2004-12-28 JP JP2006510166A patent/JP4665898B2/ja not_active Expired - Fee Related
- 2004-12-28 EP EP04807959A patent/EP1722572B1/en not_active Expired - Fee Related
- 2004-12-28 US US10/590,245 patent/US8031951B2/en not_active Expired - Fee Related
- 2004-12-28 CN CN2004800428269A patent/CN1943243B/zh not_active Expired - Fee Related
- 2004-12-28 WO PCT/JP2004/019604 patent/WO2005081539A1/ja active Search and Examination
-
2011
- 2011-09-06 US US13/226,116 patent/US8265402B2/en not_active Expired - Fee Related
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0316490A (ja) * | 1989-06-14 | 1991-01-24 | Matsushita Electric Ind Co Ltd | 動画像符号化装置 |
JPH09148938A (ja) * | 1995-11-21 | 1997-06-06 | Seiko Epson Corp | データ圧縮・伸長方法及びその装置 |
JPH1169164A (ja) * | 1997-08-21 | 1999-03-09 | Toshiba Corp | 画像符号化方法および画像符号化装置および画像復号化装置および画像形成装置 |
JPH11317950A (ja) * | 1998-04-30 | 1999-11-16 | Sony Corp | データ符号化装置、データ符号化方法及びデータ伝送方法 |
JP2000004364A (ja) * | 1998-06-15 | 2000-01-07 | Ricoh Co Ltd | 符号化/復号化装置 |
JP2000316174A (ja) * | 1999-04-30 | 2000-11-14 | Ricoh Co Ltd | カラー画像圧縮方法およびカラー画像圧縮装置 |
JP2002016808A (ja) * | 1999-05-10 | 2002-01-18 | Ricoh Co Ltd | 符号化装置、復号化装置、符号化方法、復号化方法、およびその方法を実行するためのコンピュータが読取可能な記録媒体 |
JP2001025018A (ja) * | 1999-07-06 | 2001-01-26 | Ricoh Co Ltd | ウェーブレット変換装置、符号化復号化装置、ウェーブレット変換処理方法、及び、記録媒体 |
JP2001352543A (ja) * | 2000-06-07 | 2001-12-21 | Canon Inc | 画像符号化システム及び画像符号化装置及び方法及び記憶媒体 |
JP2002232721A (ja) * | 2001-01-31 | 2002-08-16 | Canon Inc | 画像符号化装置及びその方法並びに記憶媒体 |
JP2003204439A (ja) * | 2002-01-08 | 2003-07-18 | Canon Inc | 画像符号化装置及び画像符号化方法 |
JP2003274190A (ja) * | 2002-03-12 | 2003-09-26 | Canon Inc | 静止画像符号化方式 |
JP2003274185A (ja) * | 2002-03-19 | 2003-09-26 | Sanyo Electric Co Ltd | 画像処理方法とその方法を利用可能な画像符号化装置 |
Non-Patent Citations (4)
Title |
---|
MATSUMARA S. ET AL: "Wavelet Henkan GAzo no Sub-band Kaiso. Denryoku Class-betsu Vector Ryoshikaho.", 1995 NEN GAZO FUGOKA SYMPOSIUM (PCSJ95), vol. 10, 1995, pages 121 - 122, XP002991272 * |
MATSUMURA S. ET AL: "Wavelet Henkan Gazo no Denryoku Tekio Vector Ryoshikaho.", THE TRANSACTIONS OF THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS., vol. 80-D-II, no. 1, 1997, pages 139 - 147, XP002991273 * |
OHIRA H. ET AL: "Gauss Konsei Model niyoru Gazo no Risan Wavelet Entropy Fugaka.", ITE TECHNICAL REPORT., vol. 22, no. 71, 17 December 1998 (1998-12-17), pages 37 - 42, XP002991275 * |
TOSAKA H. ET AL: "Wavelet Henkan niyoru Koseisai Seishi Gazo Asshuku no Kento.", THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS GIJUTSU KENKYU HOKOKU, DSP-96, SST96-41, CS96-59., vol. 96, no. 187, 26 July 1996 (1996-07-26), pages 31 - 36, XP002991274 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2942935A1 (en) * | 2006-07-13 | 2015-11-11 | NEC Corporation | Wavelet-based image encoding and decoding device and encoding method and decoding method |
JP2008288841A (ja) * | 2007-05-17 | 2008-11-27 | Sony Corp | 符号化装置および方法 |
JP4525704B2 (ja) * | 2007-05-17 | 2010-08-18 | ソニー株式会社 | 符号化装置および方法、記録媒体、並びにプログラム。 |
US8395634B2 (en) | 2007-05-17 | 2013-03-12 | Sony Corporation | Method and apparatus for processing information |
US8594427B2 (en) | 2007-05-17 | 2013-11-26 | Sony Corporation | Method and apparatus for reducing memory capacity in encoding image data |
US8218645B2 (en) | 2009-07-17 | 2012-07-10 | Nec Corporation | Method and apparatus for encoding and decoding by wavelet transform |
WO2011013192A1 (ja) * | 2009-07-27 | 2011-02-03 | 株式会社東芝 | 画像符号化方法および画像復号方法 |
JP5492206B2 (ja) * | 2009-07-27 | 2014-05-14 | 株式会社東芝 | 画像符号化方法および画像復号方法、ならびに、画像符号化装置および画像復号装置 |
Also Published As
Publication number | Publication date |
---|---|
CN1943243A (zh) | 2007-04-04 |
JPWO2005081539A1 (ja) | 2008-04-24 |
US8031951B2 (en) | 2011-10-04 |
EP1722572A4 (en) | 2010-04-28 |
JP4665898B2 (ja) | 2011-04-06 |
EP1722572B1 (en) | 2013-02-20 |
CN1943243B (zh) | 2013-06-05 |
US20110317929A1 (en) | 2011-12-29 |
US8265402B2 (en) | 2012-09-11 |
US20070165959A1 (en) | 2007-07-19 |
EP1722572A1 (en) | 2006-11-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP4665898B2 (ja) | 2次元信号の符号化/復号方法および装置 | |
US9635262B2 (en) | Motion estimation and detection in a camera system accelerator architecture | |
US6873343B2 (en) | Scalable graphics image drawings on multiresolution image with/without image data re-usage | |
JP4254017B2 (ja) | 画像符号化装置及び方法 | |
US20020021758A1 (en) | System and method for efficient transmission and display of image details by re-usage of compressed data | |
US7907785B2 (en) | Image processing apparatus and image processing method | |
US20030123744A1 (en) | Multi-resolution image data management system and method based on tiled wavelet-like transform and sparse data coding | |
EP1349393A1 (en) | Image compression device, image decompression device, image compression/decompression device, program for executing on a computer to perform functions of such devices, and recording medium storing such a program | |
US6978048B1 (en) | Encoding method and apparatus | |
JP4780172B2 (ja) | 情報処理装置および方法 | |
JP2008533787A (ja) | 圧縮済領域における静止画像を処理するための方法、コンピュータプログラムプロダクト、および装置 | |
US20210076066A1 (en) | Electronic apparatus and controlling method thereof | |
US7542611B2 (en) | Image processing apparatus and method for converting first code data sets into second code data for JPEG 2000 and motion JPEG 2000 | |
WO2001069585A1 (en) | System and method for efficient transmission and display of image details by re-usage of compressed data | |
KR100529540B1 (ko) | 웨이블릿 변환을 이용한 영상압축방법 | |
US11259051B2 (en) | Pyramid algorithm for video compression and video analysis | |
US20030039400A1 (en) | Hardware context vector generator for JPEG2000 block-coding | |
EP1523170A1 (en) | Image conversion device, image conversion method, and recording medium | |
JPH09327018A (ja) | 画像符号化装置及び画像復号化装置 | |
EP2787737A1 (en) | Hierarchical image and video codec | |
JP3990949B2 (ja) | 画像符号化装置及び画像符号化方法 | |
JP2000350038A (ja) | 画像データを変換する装置 | |
US20110091119A1 (en) | Coding apparatus and coding method | |
JP2004214982A (ja) | 画像処理方法 | |
JP2003234905A (ja) | 固定長画像符号化装置及び固定長画像復号化装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2006510166 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2004807959 Country of ref document: EP Ref document number: 2007165959 Country of ref document: US Ref document number: 10590245 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200480042826.9 Country of ref document: CN |
|
WWP | Wipo information: published in national office |
Ref document number: 2004807959 Country of ref document: EP |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
WWP | Wipo information: published in national office |
Ref document number: 10590245 Country of ref document: US |