WO2006009319A1 - 画像処理システム - Google Patents
画像処理システム Download PDFInfo
- Publication number
- WO2006009319A1 WO2006009319A1 PCT/JP2005/013890 JP2005013890W WO2006009319A1 WO 2006009319 A1 WO2006009319 A1 WO 2006009319A1 JP 2005013890 W JP2005013890 W JP 2005013890W WO 2006009319 A1 WO2006009319 A1 WO 2006009319A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- frequency analysis
- image
- dimensional
- fourier transform
- image processing
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1335—Combining adjacent partial images (e.g. slices) to create a composite input or reference pattern; Tracking a sweeping finger movement
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/117—Identification of persons
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/431—Frequency domain transformation; Autocorrelation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
Definitions
- the present invention relates to an image processing system, an image processing method, and an image processing program software, and more particularly, to an image processing system, an image processing method, and an image processing program / software for frequency analysis of an input image.
- Background technology :
- An image like a fingerprint ridge can be considered as a straight line with a regular interval when viewed in a narrow area. For this reason, for such an image, the image is divided into small regions that can be regarded as periodic patterns, and frequency analysis such as Fourier transform is performed for each small region. It can be carried out.
- a fingerprint / palmprint image processing system for frequency analysis of a fingerprint or palmprint image (hereinafter referred to as a finger / palmprint).
- Patent Document 1 Japanese Patent Laid-Open Publication No. Sho 5 9-2 2 1 7 3
- Patent Document 1 describes pre-processing when coding, reduction of post-processing, and simple fingerprint image coding.
- the “fingerprint image coding method” is disclosed.
- the fingerprint image encoding method described in Patent Document 1 the entire fingerprint image is divided into small regions, Fourier transform is performed for each small region, and the fingerprint component is encoded by encoding the Fourier component of the small region. ing.
- Patent Document 2 Japanese Patent Laid-Open No. 2 0 2-9 9 9 12 removes noise from an input image such as a fingerprint image and improves the image quality.
- Image processing device is disclosed.
- the block dividing unit divides an input image into blocks, and the Fourier transform unit performs a Fourier transform on the image of each block.
- the noise power suppression means estimates the power of the original image and the noise power at each frequency, sets a gain so that the larger the relative magnitude of these powers, the smaller the value, and the gain is a Fourier component. Amplitude of corresponding frequency of information Noise power is suppressed by multiplying by.
- the inverse Fourier transform means performs inverse Fourier transform on the Fourier component whose amplitude has been changed, and the image reconstruction means reads each block image subjected to inverse Fourier transform to restore the entire image.
- the image processing apparatus described in Patent Document 2 divides the entire fingerprint image into small regions, performs Fourier transform for each small region, and selects a component representing a ridge from the power of the Fourier component for each region. The fingerprint image is reconstructed using the components selected for each region.
- Patent Document 3 Japanese Laid-Open Patent Publication No. 9 1 6 7 2 30 (corresponding US Pat. No. 5, 9 3 7, 0 8 2) (hereinafter referred to as “Patent Document 3”) has properties similar to lines.
- a “finger / palm print image processing device” that correctly extracts a ridge image from a skin pattern image even in the presence of an image is disclosed.
- the local information extraction unit extracts a plurality of ridge candidate images for each local region obtained by dividing the input image.
- the first ridge candidate image selection unit selects, for each local region, a first ridge candidate image that is a ridge candidate image having a high degree of ridge-likeness.
- the connectivity evaluation unit evaluates the connectivity between the local regions of the first ridge candidate image and integrates the local regions evaluated as having good connectivity with each other in the clustering unit to generate a local region group.
- the cluster evaluation unit determines an initial local region group that is a local region group having a high degree of ridge-likeness.
- the optimum ridge candidate image selection unit selects a ridge candidate image having good connectivity with the surroundings from among the ridge candidate images in the local region that does not belong to the initial local region group.
- the ridge image restoration unit outputs the first ridge candidate image for the local region belonging to the initial local region group, and outputs the optimum ridge candidate image for the other local region as the ridge extraction result.
- Patent Document 4 Japanese Patent Laid-Open No. 2 0 2 — 2 8 8 6 4 1 (corresponding US Patent Publication No. 2 0 0 2/0 1 6 4 0 5 6) (hereinafter referred to as “Patent Document 4”)
- the “finger / palm print image processing device” is disclosed that can extract ridges more accurately even in regions where the ridges have high curvature.
- a ridge image estimated to represent a ridge is selected from among ridge candidate images for a local region other than the high-reliability region, and the selected image is selected. Each ridge image is judged to be effective as an image representing a ridge.
- Patent Document 5 An “finger and palm print image processing device” is disclosed that can extract ridges and lines more accurately even in regions where the curvature is high even in areas with high curvature of ridges.
- the region characteristic determination unit determines whether the ridge is a single region or a region where a ridge and a ridge are mixed.
- the adaptive ridge candidate selection means selects an image representing a ridge from the ridge candidate images according to the evaluation criteria corresponding to each of the ridge ridge mixed region and the ridge single existence region.
- the entire fingerprint image is divided into small regions, Fourier transform is performed for each small region, and each region is continuously connected with the Fourier component of the adjacent region.
- the component representing the ridge is selected based on the nature, and the fingerprint image is reconstructed using the component selected for each region.
- Patent Document 6 Japanese Laid-Open Patent Publication No. 10-9 976 9 (corresponding US Pat. No. 6, 2 8 9, 11 14) (hereinafter referred to as “Patent Document 6”) is sufficiently accurate and reliable. It discloses a “fingerprint reading system” that can recognize and determine the best image at a reasonable price.
- the fingerprint reading system described in Patent Document 6 reads the fingerprint when a finger and a sensor attached to the reading means are in contact with each other and the sensor and the finger slide relative to each other. And means for reconstructing a fingerprint image from the partial images obtained during the exercise.
- the finger / palm-print image processing devices of Patent Document 1 to Patent Document 5 perform a two-dimensional Fourier transform of the vertical and horizontal pixel values collectively, If the technique as disclosed in Patent Document 6 is applied and an image of the entire fingerprint or the entire small area is not constructed in advance, the two-dimensional Fourier transform process cannot be started. As a result, there is a problem that the structure of the input means is limited in addition to the redundant processing time.
- the Fourier transform must be performed in a sufficiently wide range so that the periodicity of the ridges in each region can be clearly understood.
- the image enhancement cannot be performed in detail if the encoding is performed because the image enhancement is performed for each small area. Therefore, it is possible to improve the stability of frequency analysis by ensuring the details of coding and image enhancement by performing Fourier transform on a wide area centering on the small area without making the small area larger than necessary. Is desirable.
- Patent Document 3 an example is shown in which an image of a 64 pixel square part centered on a small area is subjected to Fourier transform in order to obtain a frequency component of the small area of 8 pixel square. ing.
- the target of the Fourier transform becomes large.
- 64 pixels squarely for each small area there is a problem that the two-dimensional Fourier transform must be performed on the image of this part, and the time required for the arithmetic processing becomes extremely redundant.
- Patent Document 3 the technical idea of improving the stability of frequency analysis by guaranteeing the details of image enhancement by performing Fourier transform on a wide part centering on a small region is described in Patent Document 3.
- the present invention can be applied not only to the finger / palm-print image processing apparatus but also to any of the finger / palm-print image processing apparatuses disclosed in Patent Documents 1, 2, and 3 to 5, for example.
- problems such as the standby time for image input and the redundancy of the time required for computation processing have become a bottleneck.
- Patent Document 7 Special Table 2 0 0 1—5 1 1 5 6 9 (corresponding US Pat. No. 5, 9 5 3, 4 4 2) (hereinafter referred to as “Patent Document 7”) is an unknown fingerprint.
- a method and apparatus for automatically positioning a first unknown image, such as an image, to one of a plurality of force categories is disclosed.
- the invention disclosed in Patent Document 7 is derived from the frequency display of each image category. Storing a plurality of value sets in the library.
- the classification processing method and apparatus takes a frequency image of a first unknown pattern and creates a first frequency image.
- the frequency image plane of the first (unknown) frequency image is divided into a plurality of frequency image plane regions. Each frequency image plane region may be a corner segment extending radially from the origin of the frequency image plane.
- a region value is assigned to each frequency image plane region based on the total energy in the frequency image within the angular frequency image plane region.
- the region values for the first frequency image are combined to generate a first region value set.
- the first region value set is compared in a comparator with each of the stored value sets.
- the comparator preferably performs a correlation function on the set or pattern of region values using a one-dimensional frequency transform of the spatial representation of the pattern of region value sets.
- Patent Document 7 the pixel values of a fingerprint image are collectively subjected to a two-dimensional Fourier transform. Accordingly, even in the methods and apparatus disclosed in Patent Document 7, there same problems force s and palm print image processing device as disclosed in Patent Documents 1 to 5 described above.
- an object of the present invention is to provide an image processing system, an image processing method, and an image processing program capable of reducing the overall processing time required for frequency analysis processing. Disclosure of the invention:
- An image processing system is an image processing system that divides an original image into small regions and performs frequency analysis for each small region. Necessary for partial input image input device, first frequency analysis unit that divides partially captured image into pixel columns, and performs one-dimensional frequency analysis of each pixel column, and small area analysis And a second frequency analysis unit that collects the one-dimensional frequency analysis results obtained by the first frequency analysis unit and obtains the two-dimensional frequency analysis results in a small area.
- an image input device such as a finger / palm print sensor partially captures an image such as a fingerprint or palm print, and the first frequency analysis unit force is partially captured.
- the image is divided into pixel rows, and each pixel row is subjected to one-dimensional frequency analysis to obtain a one-dimensional frequency analysis result.
- the i-th frequency analysis unit can start one-dimensional frequency analysis when at least one column of pixels has been captured via the image input device, reducing the waiting time for image input. Is done.
- the first frequency analysis unit is a new image that is partially captured via the image input device until the number of one-dimensional frequency analysis results necessary to perform small-area analysis is obtained.
- One-dimensional frequency analysis is repeated for (at least one column of pixels), and one-dimensional frequency analysis results are obtained one after another.
- the second frequency analysis unit S collects the one-dimensional frequency analysis results, and based on these one-dimensional frequency analysis results, Obtain 2D frequency analysis results of the region.
- the first frequency analysis unit S shortens the waiting time until the start of the one-dimensional frequency analysis, and the first frequency analysis unit preliminarily parallels the image capture to The number of one-dimensional frequency analysis results required to perform the analysis is obtained, and the second frequency analysis unit obtains the two-dimensional frequency analysis results in a small area based on these one-dimensional frequency analysis results.
- the overall processing time required for the frequency analysis processing can be reduced.
- An image processing system includes an image input device that sequentially captures an original image sequentially, divides the partially captured image into pixel columns, and a first-dimensional frequency analysis of each pixel column.
- 2nd frequency analysis module that collects 1D frequency analysis results obtained by 1 frequency analysis unit and 1st frequency analysis unit necessary for small area analysis to obtain 2D frequency analysis results And a configuration provided with a base.
- the overall configuration is the same as that described above except that the function of dividing the partially captured image into pixel rows is provided on the image input device side instead of the first frequency analysis unit.
- An image processing system sequentially captures an original image sequentially,
- An image input device having a first frequency analysis unit that divides a partially captured image into pixel columns and performs one-dimensional frequency analysis on each pixel column, and a second region necessary for analyzing a small region. It has a configuration with a second frequency analysis unit that collects the 1D frequency analysis results obtained by the 1 frequency analysis unit and obtains the 2D frequency analysis results.
- the image input device side has the function of dividing the partially captured image into pixel rows and the function of the first frequency analysis unit for one-dimensional frequency analysis of each divided pixel row.
- the overall configuration is the same as described above.
- the first frequency analysis unit is assumed to perform one-dimensional Fourier transformation on the pixel column
- the second frequency analysis unit is assumed to be one-dimensional Fourier transformation obtained by the first frequency analysis unit.
- the results can be collected and further subjected to a one-dimensional Fourier transform to obtain a two-dimensional Fourier transform result of the original image.
- the second frequency analysis unit collects these one-dimensional Fourier transform results and further performs one-dimensional Fourier transform to obtain the two-dimensional Fourier transform result of the original image.
- the frequency component for each small area is obtained by repeatedly extracting the small area continuously.
- the frequency analysis results obtained previously by the first frequency analysis unit are used repeatedly, so that calculations related to 1D frequency analysis of overlapping pixel arrays are performed. Can be omitted to speed up the processing.
- This image processing system can be used in an image processing system that captures a fingerprint or palm print image from an image input device, that is, a finger / palm print image processing system. This provides an image processing system for a finger / palm-print with excellent processing speed.
- this image processing system may be used for an image processing system for a finger / palm print that detects the vicinity of a feature point of a fingerprint or palm print from the frequency analysis result and restores only the ridge near the feature point.
- an image processing system for a finger / palm-print that can restore the ridge near the feature point in a short time is provided.
- the first frequency analysis unit the second frequency analysis unit, or the first and second frequency analysis units, it is possible to speed up the processing by performing frequency analysis using a microprocessor dedicated to signal processing. is there.
- each of the first and second frequency analysis units with a dedicated microprocessor for signal processing, some processing such as image capture, segmentation, one-dimensional frequency analysis, and two-dimensional frequency analysis can be performed. It is possible to execute in parallel, and the overall processing speed is increased.
- the image processing method is an image processing method that divides an original image into small regions and performs frequency analysis for each small region.
- first frequency analysis processing that divides each partially captured image into pixel rows and performs one-dimensional frequency analysis on each pixel row, and analysis of small regions
- one-dimensional frequency analysis is started to shorten the waiting time, and the number of one-dimensional frequency analysis results required to perform small area analysis Is obtained in advance in parallel with image capture, and the processing load is reduced by obtaining the two-dimensional frequency analysis result of a small region based on these one-dimensional frequency analysis results. Reduced overall processing time required Appear.
- the image processing method includes an image input process in which an original image is sequentially partially captured, and the partially captured image is divided into pixel columns, and each pixel column is subjected to one-dimensional frequency analysis. 2nd frequency to collect 2D frequency analysis results by collecting 1D frequency analysis results obtained by 1st frequency analysis processing and 1st frequency analysis processing necessary for small area analysis And an analysis process to frequency-analyze the image.
- the image processing method includes a first frequency that sequentially captures an original image, divides the partially captured image into pixel columns, and performs one-dimensional frequency analysis on each pixel column.
- the second frequency analysis that collects the 1D frequency analysis results obtained by the image input processing for analysis processing and the first frequency analysis processing necessary for small area analysis to obtain the 2D frequency analysis results And processing to perform frequency analysis of the image.
- the image input process includes the process of dividing the partially captured image into pixel columns and the first frequency analysis process that performs one-dimensional frequency analysis of each divided pixel column.
- the configuration is the same as described above.
- the first frequency analysis process is a one-dimensional Fourier transform of a pixel row
- the second frequency analysis process is a collection of the one-dimensional Fourier transform results obtained in the first frequency analysis process. Furthermore, it can be configured to obtain a 2D Fourier transform result of the original image by performing a 1D Fourier transform.
- the one-dimensional Fourier transform of the pixel columns is started immediately to shorten the waiting time and to analyze the small area.
- the number of 1D Fourier transform results is obtained in advance, and the processing load is reduced by obtaining the 2D Fourier transform results of the original image based on these 1D Fourier transform results.
- the overall processing time required for the process is shortened.
- a small region is repeatedly extracted repeatedly to obtain a frequency component for each small region.
- the second and subsequent two-dimensional frequency angle analysis in the second frequency analysis process repeatedly uses the frequency analysis result previously obtained in the first frequency analysis process, thereby overlapping pixels. It is possible to speed up the processing by omitting calculations related to the one-dimensional frequency analysis of the columns.
- the first frequency analysis process executed before or before is obtained.
- the frequency analysis result overlaps can be used as they are, so the number of executions of the first frequency analysis process required to execute the second and subsequent 2D frequency analysis processes is reduced, and the frequency Significant reduction in overall processing time required for analysis processing is realized.
- This image processing method can be used as an image processing method for capturing a fingerprint or palm print image by image input processing, that is, an image processing method for a finger / palm print. As a result of shortening the processing time required for two-dimensional frequency analysis of the original image, the overall processing speed of image processing for the palm and palm print is reduced.
- this image processing method can be used for an image processing method that detects the vicinity of a feature point of a fingerprint or palm print from the frequency analysis result and restores only the ridge near the feature point.
- the processing time required for the two-dimensional frequency analysis of the original image is shortened.
- the ridge near the feature point can be restored in a short time.
- the image processing program software divides the original image into small regions and sequentially transmits the original images to the microphone mouth processor provided in the image processing system that performs frequency analysis for each small region.
- a partial image capture control function a partially captured image divided into pixel columns, a first frequency analysis function that analyzes each pixel column in a one-dimensional frequency, and a small area analysis This is to realize a second frequency analysis function that collects the one-dimensional frequency analysis results obtained in the first frequency analysis necessary for obtaining a two-dimensional frequency analysis result in a small area.
- one-dimensional frequency analysis is started to shorten the waiting time, and the number of one-dimensional frequency analysis results required to perform small area analysis These are obtained in advance in parallel with image capture, and these 1
- the overall processing time required for frequency analysis processing can be reduced.
- the image processing program software divides the original image into small areas and sequentially sends the original image to the microphone mouth processor installed in the image processing system that performs frequency analysis for each small area. Necessary for analyzing a small area, an image input control function that divides partially captured images and divides the partially captured image into pixel columns, a first frequency analysis function that performs one-dimensional frequency analysis of each pixel column The second frequency analysis function that collects the 1D frequency analysis results obtained by the first frequency analysis and obtains the 2D frequency analysis results is realized.
- the overall configuration is the same as described above, except that the function of dividing the partially captured image into pixel columns is included in the image input control function instead of the first frequency analysis function.
- the image processing program software according to the third aspect of the present invention divides the original image into small areas and sequentially converts the original image into a microphone port processor provided in an image processing system that performs frequency analysis for each small area.
- First frequency analysis function that divides a partially captured image into pixel columns and performs one-dimensional frequency analysis of each pixel column, and first frequency analysis necessary for small area analysis This is to realize the second frequency analysis function that collects the 1D frequency analysis results obtained in step 2 and obtains the 2D frequency analysis results.
- the first frequency analysis function is a one-dimensional Fourier transform of a pixel row
- the second frequency analysis function is a collection of the one-dimensional Fourier transform results obtained in the first frequency analysis. Furthermore, it can be configured to obtain a two-dimensional Fourier transform result of the original image by performing a one-dimensional Fourier transform.
- the one-dimensional Fourier transform of the pixel columns is started immediately to shorten the waiting time, and the analysis of the small area is performed.
- the number of one-dimensional Fourier transform results required to be obtained is obtained in advance, and the processing load is reduced by obtaining the two-dimensional Fourier transform results of the original image based on these one-dimensional Fourier transform results.
- the overall processing time required for the 2D Fourier transform process is reduced.
- the frequency component for each small area is obtained by repeatedly extracting the small area continuously.
- the frequency analysis result obtained previously by the first frequency analysis function is repeatedly used, so that the one-dimensional frequency of the overlapping pixel array is Calculations related to analysis can be omitted to speed up processing.
- This image processing program software can be used for image processing program software that captures fingerprint or palm print images with the image input control function, that is, image processing program software for finger palm prints. As a result of the reduction in processing time required for 2D frequency analysis of the original image, the overall processing speed of image processing for the palm and palm print is reduced.
- This image processing program software can also be used for image processing program software that detects the vicinity of feature points of fingerprints or palm prints from frequency analysis results and restores only the ridges near the feature points. Good. As described above, as a result of shortening the processing time required for the two-dimensional frequency analysis of the original image, the ridges near the feature points can be restored in a short time.
- the image processing system, image processing method, and image processing program software of the present invention starts the one-dimensional frequency analysis immediately when the original image is partially captured and divided into pixel columns. Wait for image input until analysis starts The machine time can be shortened.
- the image processing system, image processing method, and image processing program software of the present invention obtains the result of the one-dimensional frequency analysis in advance almost in parallel with the acquisition of the original image, and uses the result of the one-dimensional frequency analysis. Since the processing load is reduced by obtaining the two-dimensional frequency analysis result of the small area based on the conventional method, the vertical and horizontal pixel values of the small area are collectively subjected to two-dimensional Fourier transform for frequency analysis. Compared with, the overall processing time required for frequency analysis processing can be greatly reduced.
- FIG. 1 is a functional block diagram showing a simplified image processing system according to the first embodiment of the present invention.
- FIG. 2 is a block diagram showing an image input device used in the image processing system shown in FIG.
- FIG. 3 is a functional block diagram showing a simplified computer that can be used in the image processing system shown in FIG.
- FIG. 4 is a flowchart showing the overall operation of the image processing system of the embodiment.
- FIG. 5 is a diagram showing an example of the result of restoring the ridge only in the vicinity of the feature point.
- FIG. 6 is a conceptual diagram showing an example of small area division.
- Figure 7 is a conceptual diagram showing an example of the relationship between a small region and the frequency analysis target (Fourier transform region).
- Fig. 8 is a conceptual diagram showing the relationship between the one-dimensional Fourier transform and the two-dimensional Fourier transform.
- Fig. 9 is a conceptual diagram showing an example of overlapping states of multiple frequency analysis objects (Fourier transform regions).
- FIG. 10 is a flowchart specifically showing an example of the processing operation of the microprocessor of the computer functioning as the first and second frequency analysis units.
- Fig. 11 is an operational principle diagram showing an example of reducing the processing load by reusing the one-dimensional frequency analysis results obtained by the first frequency analysis unit.
- FIG. 12 shows a simplified image processing system according to the second embodiment of the present invention. It is a functional block diagram.
- FIG. 13 is a block diagram showing an image input device used in the image processing system shown in FIG.
- FIG. 14 is a functional block diagram showing a simplified computer used in the image processing system shown in FIG.
- FIG. 15 is a flowchart showing the overall operation of the image processing system according to the embodiment.
- FIG. 16 is a functional block diagram schematically showing an image processing system according to the third embodiment of the present invention.
- FIG. 17 is a block diagram showing an image input device used in the image processing system shown in FIG.
- FIG. 18 is a functional block diagram showing a simplified computer used in the image processing system shown in FIG.
- FIG. 19 is a flowchart showing the overall operation of the image processing system according to the embodiment.
- FIG. 20 is a functional block diagram schematically showing an image processing system according to the fourth embodiment of the present invention. Best Mode for Carrying Out the Invention:
- FIG. 1 is a functional block diagram schematically showing the image processing system 20 according to the first embodiment of the present invention.
- the image processing system 20 is a system that divides an original image into small areas and performs frequency analysis for each small area.
- the illustrated image processing system 20 is roughly composed of an image input device 30 such as an image sensor, a computer (CPU) 40 0 operated by program control, and an operation result output device 50 such as a display. ing.
- the illustrated computer 40 includes a first frequency analysis unit 42, a second frequency analysis unit 44, and a feature line detection Z ridge restoration unit 46.
- the image input device 30 includes an image capturing unit 62.
- the image capturing unit 62 sequentially captures the original image sequentially and outputs the partially captured image to the computer 40 sequentially.
- the original image may be a fingerprint or palm print.
- the first frequency analysis unit 42 includes an image dividing unit 64 and a first frequency analysis unit 66.
- the second frequency analysis unit 44 has a second frequency analysis unit 68.
- the image dividing unit 64 divides the partially captured image into pixel columns and sequentially outputs the pixel columns.
- the first frequency analysis unit 66 performs one-dimensional frequency analysis on each pixel column and sequentially obtains a one-dimensional frequency analysis result.
- the first frequency analysis unit 42 sequentially captures pixel columns having a predetermined width from the image input device 100 and performs one-dimensional frequency analysis on each pixel column.
- the second frequency analysis unit 68 collects as many 1-dimensional frequency analysis results as necessary to perform the analysis of the small region, and obtains the two-dimensional frequency analysis result of the small region.
- the second frequency analysis unit 44 collects a predetermined number of one-dimensional frequency analysis results obtained by the first frequency analysis unit 42, and re-analyzes these one-dimensional frequency analysis results. 2D frequency analysis results are obtained by performing 1D frequency analysis.
- the feature point detection Z ridge restoration unit 4 6 detects the vicinity of the feature point of the fingerprint or palm print from the two-dimensional frequency analysis result, and restores only the ridge near the feature point. The result is output to the calculation result output device 1 2 0.
- an original image is input from the image input device 30 (step A l).
- the image capturing unit 62 of the image input device 30 sequentially captures the original image sequentially and transmits the captured part (partially captured image) to the computer 40 as needed (step A 2).
- the image dividing unit 64 of the first frequency analysis unit 42 divides the transmitted image (partially captured image) into a pixel row having a predetermined width (step A 3).
- the first frequency analysis unit 66 of the first frequency analysis unit 42 performs one-dimensional frequency analysis for each divided pixel column, and the one-dimensional frequency analysis result of each pixel column. (Step A4).
- the second frequency analysis unit 68 of the second frequency analysis unit 44 has a one-dimensional value calculated for each pixel column by the first frequency analysis unit 66 of the first frequency analysis unit 42. A predetermined number of frequency analysis results are collected, and the one-dimensional frequency analysis is performed again on these one-dimensional frequency analysis results to obtain a two-dimensional frequency analysis result (step A 5).
- the feature point detection / ridge restoration unit 46 first detects the vicinity of the feature point of the fingerprint or palm print from the two-dimensional frequency analysis result (step A 6). Subsequently, the feature point detection / ridge restoration unit 46 restores only the ridge near the feature point (step A 7). Then, the calculation result output device 50 displays this ridge restoration result (step A 8).
- Figure 5 shows an example of the result of restoring the ridges only near the feature points.
- the image processing system 20 according to the first embodiment of the present invention is configured to perform frequency analysis for each pixel column having a predetermined width, the frequency analysis processing is started before the entire image is input. it can.
- the 1D frequency analysis results obtained by the 1st frequency analysis unit 4 2 are reproduced by the 2D frequency analysis process in the 2nd frequency analysis unit 4 4. By using it, the processing time can be shortened.
- each device unit
- this image processing system 20 is used as a finger / palm-print image processing system.
- a finger / palm print sensor is used as the image input device 30, and a personal computer is used as the computer 40, and the personal computer 40 has a display functioning as the calculation result output device 50.
- the personal computer 4 ⁇ has a first frequency analysis unit 42, a second frequency analysis unit 44, and a feature point detection Z ridge restoration unit 46.
- an image is input to the personal computer 40 from the finger / palm print sensor 30 via the network, or a finger already recorded in a storage device such as a memory or a hard disk inside or outside the personal computer 40. It is also possible to perform frequency analysis processing on palmprint images. When frequency analysis processing is performed on a finger / palm print image already recorded in a storage device such as a memory or hard disk, a finger / palm print sensor is not required for the image processing system 20 itself.
- the image input device 3 0 is constituted by a data reading device that reads a partial image from a finger / palm-print image stored in a memory or a hard disk.
- the image input from the finger / palm print sensor 30 is subjected to frequency analysis for each small area.
- the small area can be divided arbitrarily, but here it is divided into a square lattice as shown in Fig. 6.
- the sensor 30 is assumed to be 20 pixels per mm and is divided into a square lattice of 8 pixels on each side. However, other sizes and shapes may be used.
- the frequency analysis can be performed on the small area itself, or it can be performed on the surrounding image including the small area in order to evaluate the periodicity stably.
- a description will be given using an example of frequency analysis of a square part of 32 pixels per side centered on a small area. Of course, other sizes are also acceptable.
- the frequency analysis range has a size that includes at least about two ridges.
- the frequency analysis range is too large, it is strongly affected by the ridges being bent, so it is desirable to make it smaller than the size that contains about four ridges.
- the image When one column of image is input from the finger / palm print sensor 30, the image is cut out every 8 pixels (small area size) with a width of 3 2 pixels (size subject to frequency analysis).
- One-dimensional Fourier transform is performed by frequency analysis unit 4 of 1.
- the first frequency analysis unit 42 operates as a first Fourier transformer that sequentially performs a one-dimensional Fourier analysis result as a one-dimensional frequency analysis result by performing a one-dimensional Fourier transform on each divided pixel column. To do.
- the first frequency analysis unit 4 2 can operate if there is one input image, the image There is no need to wait for the entire input, and the first frequency analysis unit 42 can be activated as soon as a column of pixels is repelled.
- the second frequency analysis unit 44 performs the one-dimensional Fourier transform for 32 pieces.
- the results are collected and subjected to Fourier transform, and as shown in Fig. 8, the result of two-dimensional Fourier transform for a square image with 32 pixels per side is obtained. That is, the second frequency analysis unit 44 collects the one-dimensional Fourier transform results, further performs one-dimensional Fourier transform, and obtains a two-dimensional Fourier transform result as a two-dimensional Fourier transform result of the small region.
- the second frequency analysis unit 44 collects the one-dimensional Fourier transform results, further performs one-dimensional Fourier transform, and obtains a two-dimensional Fourier transform result as a two-dimensional Fourier transform result of the small region.
- the pixel position is defined in a coordinate system such that the origin is the lower left position of a square with 32 pixels on one side shown in Fig. 8, the right is + X, and the upper is + Y.
- the second frequency analysis unit 44 can operate, so there is no need to wait for the input of the entire image.
- a window function is often used to reduce the effect of discontinuities near the boundaries of the data. In such cases, the same processing can be performed. For example, using a Gaussian window with variance ⁇ ,
- Sequential calculation is possible by calculating. Any window function can be used as well as a Gaussian window, such as a rectangular window, which can separate calculations vertically and horizontally.
- the Fourier transform used in the second frequency analysis unit 44 In a small region that is adjacent to the direction in which the first frequency analysis unit 42 performs the one-dimensional Fourier transform, the Fourier transform used in the second frequency analysis unit 44, as shown by the hatched portion in FIG.
- the target area overlaps.
- the frequency analysis of the small area 2 in Fig. 9 is performed by the second frequency analysis unit 44
- the one-dimensional Fourier transform result of the hatched area used for the frequency analysis of the small area 1 in Fig. 9 is reused. can do. For this reason, the calculation is performed more than when performing two-dimensional Fourier transform on subregion 1 and subregion 2, respectively. The amount can be reduced.
- the period of ridges of fingerprints or palm prints is limited to a specific range, high-frequency components that cannot be used as fingerprints or ridges of palm prints are included in the first and second frequency analysis units 4. 2, 4 4 can be removed. By doing so, the number of processing objects is reduced and the amount of calculation can be reduced.
- the second frequency analysis unit 44 obtains the 2D Fourier transform result for each small area, and when the feature line detection / ridge restoration unit 46 restores the ridge, it becomes the display 50.
- the ridge restoration result is output.
- the second frequency analysis unit 44 outputs frequency information for each small region as a two-dimensional frequency analysis result.
- Feature point detection / ridge restoration unit 46 restores the fingerprint or palm print ridge from this frequency information and outputs the ridge information.
- the feature point detection Z ridge restoration unit 46 is obtained from the frequency information of each small region by using the method of Patent Document 2 to Patent Document 5 described in the background section. You may restore the ridges. In this case, the ridge information can be output correctly even if there is a hail or noise. More specifically, as described in Patent Document 2 to Patent Document 5, the feature point detection / ridge restoration unit 46 removes the noise component from the two-dimensional frequency analysis result, and inversely converts it. Also, the ridge can be restored by representing one point in the two-dimensional frequency space from the two-dimensional frequency analysis result and inversely transforming that one point.
- the inverse transform is equivalent to a two-dimensional sine wave if it is a Fourier transform.
- the feature point detection Z ridge restoration unit 46 can detect fingerprint or palm print feature points (ridge breaks or branch points) depending on the frequency information of each small region and the frequency information between each small region. It is possible to detect a probable location and output only the frequency information of that part, or restore only the ridges of that part.
- a method for determining whether or not it is near a feature point the method described in claim 6 of Japanese Patent Application No. 2 0 0 4-6 1 2 92 (hereinafter referred to as a feature point detection method) is used. Can be used.
- the feature point detection / ridge restoration unit 46 uses this feature point detection method to detect the vicinity of a fingerprint or palmprint feature point from the two-dimensional frequency analysis result, or to detect the two-dimensional frequency of adjacent small regions. By comparing with the analysis result, it is judged whether it is near the feature point or not.
- the reason for the comparison This is because the ridge flow is discontinuous with the adjacent region near the feature point, and the two-dimensional frequency analysis result is different from the two-dimensional frequency analysis result of the adjacent small region.
- the microprocessor of the personal computer 40 (hereinafter referred to as CPU) in the portion functioning as the first and second frequency analysis units 42 and 44.
- CPU the microprocessor of the personal computer 40
- the finger / palm print sensor 30 starts to operate autonomously when it receives an image capture command from the CPU of the personal computer 40, and each pixel in a square image area having 32 pixels on one side as shown in FIG. , (0, 0), (1, 0), (2, 0), ..., (31, 0), ..., (0, i), (1 , i), (2,, ..., (31, ⁇ ), ..., (0, 31), (1, 31), (2, 31), ..., (31, 31) )
- the output is serially from the lower left to the upper right.
- the CPU 40 When the finger / palm-print image processing system 20 is activated, the CPU 40 first sets the flag F to store the start of image segmentation for the first small area (step a 1) and is used in the first frequency analysis process.
- a register for specifying the pixel value storage register that stores the pixel value to be stored and the one-dimensional Fourier transform result storage register that stores the result of the one-dimensional Fourier transform (first frequency analysis process) using the pixel values for one column The values of the specific indices X and y are both initialized to 0 (step a 2), an image capture command is output to the finger / palm print sensor 30 (step a 3), and image capture by the finger / palm print sensor 30 is started.
- the CPU 40 converts the pixel value into a pixel value storage register f (X, y) (step a5), and whether or not the current value of the register specific index X has reached the set value 31, that is, input of all 32 pixel values arranged in a row in the horizontal direction is completed. (Step a6). If the input of pixel values for one column is not completed, the CPU 40 increments the value of the register specific index X by 1 (step a 7) and then prepares for the input of the next pixel value (step a 4 ).
- the pixel value storage registers f (0, 0) to f (3 1 and 0), pixel values of pixels of (0, 0) spot to (31, 0) spot in a 32 pixel square image are stored in a one-to-one correspondence as shown in FIG. Become.
- the first frequency analysis is performed.
- the CPU 40 functioning as the unit 42 is based on the 32 pixel values stored in the pixel value storage registers f (0, 0) to f (31, 0). Execute the first frequency analysis process, find the result of one-dimensional Fourier transform of the pixel row where the Y coordinate is 0, and calculate the result based on the current value of the register specific index y Store in the result storage register (p, y) (step a8).
- the CPU 40 determines whether or not the flag F is set, that is, whether or not the current one-dimensional Fourier transform process is performed during image segmentation for the first small area ( Step a 9).
- the flag F is set, which means that the current one-dimensional Fourier transform process is being performed during image segmentation for the first small area. Determines whether the current value of the specific index y has reached the set value 31, that is, whether or not the results of one-dimensional Fourier transform have been obtained for all 32 pixel columns arranged in the vertical direction. (Step a10 ).
- step a 1 the CPU 40 increments the value of the register specific index y by 1 (step a 1 1), and After the value of the register specific index X is initialized to 0 again (step a 12), the image input processing for the next one column of pixel values is started in the same manner as described above (step a 4 to step a 7). ).
- the results of the one-dimensional Fourier transform for 32 columns which is the number of one-dimensional frequency analysis results necessary for small area analysis, are obtained.
- the CPU 40 resets the flag F so that the image of the first small region is displayed.
- the fact that the cutout has been completed is stored (step a 1 3).
- the CPU 40 functioning as the second frequency analysis unit 44 1-dimensional Fourier transform result storage registers (p, 0) to Fi (p, 3 1) for 32 columns stored in the 1-dimensional Fourier transform result storage registers
- formula 2 or formula 5
- F 2 (p, q) which is the result of the two-dimensional Fourier transform of this small region, is obtained (step a 14), and output the calculation result to a display, for example (step a 15).
- the CPU 40 reuses some of the one-dimensional frequency analysis results obtained by the first frequency analysis unit 42 in the next two-dimensional Fourier transform process in order to execute the present two-dimensional Fourier transform process. Therefore, the values of the one-dimensional Fourier transform result storage registers F i (p, 0) to F ⁇ ( ⁇ '23 ) are stored in the backup registers (, 0) (p, 2
- step a 16 The process of storing in a one-to-one correspondence is executed (step a 16).
- an image is cut out every 8 pixels with a width of 32 pixels. Therefore, as is clear from FIG. 9, the image is cut out centering on the small region 1 in this processing.
- the area to be clipped is moved downward by 8 pixels, and the image is clipped around the small area 2. Therefore, among the pixels from the 0th column to the 31st column in the next Fourier transform region, the pixels from the 8th column to the 31st column should overlap with the 0th column to the 23rd column of the current Fourier transform region. become.
- step a 17 the CPU 40 determines whether or not to continue the two-dimensional frequency analysis for the next small area. If continuation is set, CPU4 0 moves the small area to be analyzed downward by 8 pixels. For example, after moving the small area to be analyzed from small area 1 to small area 2 in FIG. 9 (step a 18), The processing after step a2 is repeatedly executed in the same manner as described above for a new small area, for example, small area 2. To set or cancel the continuation of 2D frequency analysis, turn on the execution flag with a command from the keyboard of the personal computer 40.
- step a 9 the result of step a 9 is always false.
- the CPU 40 determines whether or not the current value of the register specific index y has reached the set value 7 (step a 19), and performs a loop process similar to the above to each column related to the 0th column to the 7th column.
- the one-dimensional Fourier transform results are obtained, and these results are overwritten in a one-to-one correspondence with each of the one-dimensional Fourier transform result storage registers (, 0) to Fi (p, 7) and stored.
- step a 19 When it is confirmed in step a 19 that the current value of the register specific index y has reached the set value 7, the CPU 40 functioning as the second frequency analysis unit 44 uses the backup register ( ⁇ , 0 ) To G 1 (p, 23) are overwritten in one-to-one correspondence and stored in one-dimensional Fourier transform result storage registers F 1 (p, 8) to Fi (p, 31) (step a 20 ). Then, as described above, the CPU 40, as a result, stores the one-dimensional Fourier transform results for 32 columns stored in the one-dimensional Fourier transform result storage registers (p, 0) to F (p, 31) at the present time, that is, newly I was asked
- the CPU 40 Based on the values of (p, 0) to Fi (p, 7) and the values of (p, 8) to Fi (p, 31) already obtained in the previous processing, An arithmetic expression, that is, the second frequency analysis process is executed.
- the CPU 40 then obtains the result F 2 of the two-dimensional Fourier transform of the small area that is the object of analysis at this time, for example, the small area 2 shown in FIG. Find (p, q) (step a14). Subsequently, the CPU 40 outputs the calculation result to, for example, a display (step a 15), and then executes the one-dimensional frequency analysis result obtained by the first frequency analysis unit 42 in the same manner as described above.
- the pack-up process for re-use in the two-dimensional file conversion process is executed (step a 16).
- FIG. 11 specifically shows an example of reuse of the one-dimensional frequency analysis result obtained by the first frequency analysis unit 42.
- the first frequency analysis unit 42 performs calculation processing for 32 columns (Equation 1 or Equation 4 only when performing the two-dimensional frequency analysis for the first small region. ) To store the values in the one-dimensional Fourier transform result storage registers D (p, 0) to F 1 (p, 31). In the 2nd and subsequent 2D frequency analysis processing, the 1D frequency analysis unit 42 previously calculated using the knock-up registers (p, 0) to Gi (p, 23) as an intermediary. The value of the Fourier transform result is reused repeatedly. As a result, the calculations related to the one-dimensional frequency analysis for 24 columns, which are overlapping pixel columns (Equation 1 or 4), can be omitted. In practice, the first frequency analysis unit 42 can be It can be seen that it is only necessary to execute the calculation processing for the 8 columns related to the pixels in the 7th column.
- the image processing system 20 A according to the second embodiment of the present invention is compared with the image processing system 20 OA according to the first embodiment of the present invention shown in FIG. And the operation of the first frequency analysis unit 42 A in the computer 4 OA are different.
- the image input device 30 A includes an image capturing unit 62 and an image segmentation unit. It consists of part 6 and 4.
- the first frequency analysis unit 4 2 A is composed of only the first frequency analysis unit 6 6.
- the image input device 3 O A sequentially sends a pixel array having a predetermined width from the input image (original image) to the computer 4 O A.
- the first frequency analysis unit 4 2 A is configured to perform one-dimensional frequency analysis on the pixel row sent from the image input device 30 A.
- the image capturing unit 6 2 of the image input device 3 O A inputs the original image (step B 1). That is, the image capturing unit 62 sequentially captures the original image sequentially and outputs the partially captured images sequentially.
- the image dividing unit 6 4 of the image input device 3 OA divides the partially captured image into pixel columns having a predetermined width (step B 2), and the divided pixel columns are divided. Sequentially sent to computer 40 A (step B 3). For example, when the image input device 3 OA is configured by a finger / palm print image input device, the finger / palm print image input device sequentially divides the image (original image) input from the sensor into a pixel array of 32 pixels wide.
- Computer 4 Send to OA.
- the pixel rows are sequentially transmitted from the end of the sensor and divided into pixel rows for transmission, so that the pixel rows can be sent sequentially before the entire surface is scanned.
- the pixel columns can be sequentially transmitted to the computer 4 O A before the sweep of the entire finger is completed.
- the first frequency analysis unit 62 of the first frequency analysis unit 4 2 A provided on the computer 4 OA side sequentially performs one-dimensional frequency analysis on the pixel row sent from the image input device 3 OA ( Step B 4).
- the pixel sequence is sequentially transmitted to the computer 40 A.
- the pixel sequence is temporarily transmitted to the finger / palm-print image input device 30 A.
- a buffer (not shown) is provided to store the pixel row in advance, and the pixel row required for the computer 4 OA side is requested from the finger / palm print image input device 3 OA, and the finger / palm print image is obtained. It is also possible for the input device 3 OA to read out the requested pixel column from the buffer and transmit it to the computer 40 A.
- the operations of the operation result output device 50 in FIG. 8 are the operations of the second frequency analysis unit 44, the feature point detection / ridge restoration unit 46 in the embodiment of FIG. 1, and the operation result output device 50, respectively.
- Step A 5 in FIG. 4, Steps A 6 and A 7 and Step A 8) are the same as those in FIG.
- the image input device 3 OA such as a finger / palm-print image input device divides the input image (original image) into pixel columns and sequentially transmits them to the computer 4 OA. Therefore, the pixel columns are transmitted to the computer 4 OA.
- the first frequency analysis unit 4 2 A can be activated and the process can be started before the entire image is input.
- FIGS. 1 to 4 the overall configuration is shown in FIGS. 1 to 4 except that the function of dividing the pixel row is omitted from the first frequency analysis unit 42 A and provided on the image input device 30 A side. This is substantially the same as the case of the first embodiment of the present invention described.
- the image processing system 20 B according to the third embodiment of the present invention is different from the image processing system 20 A shown in FIG.
- the operation of 0 B is different as will be described later.
- the image input device 30 B includes an image capturing unit 62, an image dividing unit 64, and a first frequency analysis unit 42B.
- the first frequency analysis unit 4 2 B has a first frequency analysis unit 6 6.
- the computer 40 B includes a second frequency analysis unit 44 and a feature point detection ridge restoration unit 46.
- the second frequency analysis unit 44 has a second frequency analysis unit 68.
- the image input device 30 B of the image processing system 20 B has a first frequency analysis unit 4 2 B in addition to the function of dividing the pixel column,
- the first frequency analysis unit is omitted from the computer 4 OB side.
- the image input device 30 B extracts a pixel row having a predetermined width from the input image (original image) force and performs a one-dimensional frequency analysis by the first frequency analysis unit 4 2 B.
- the result of the one-dimensional frequency analysis is sequentially sent to the computer 40 B.
- the second frequency analysis unit 44 obtains the two-dimensional frequency analysis result, and this two-dimensional frequency analysis result is obtained.
- the feature point detection / ridge restoration unit 46 restores the ridge from the analysis result, and outputs the ridge restoration result to the calculation result output device 50.
- the image capturing unit 62 of the image input device 30 B sequentially captures the original image sequentially and sequentially outputs the partially captured image (step C 1).
- the 30 B image dividing unit 64 divides the partially captured image into pixel columns having a predetermined width, and sequentially outputs the pixel columns (step C 2).
- the first frequency analysis unit 4 2 B in the image input device 30 B performs a one-dimensional frequency analysis on the pixel array (step C 3), and performs a one-dimensional frequency analysis. The results are sequentially sent to the computer 40 B (step C 4).
- the finger / palm-print image input device 30 B sequentially divides an image input from the sensor into a pixel row having a width of 32 pixels, and Palmprint image input device 30 B
- the first frequency analysis unit provided in B itself 4 2 B performs a one-dimensional Fourier transform of the pixel array, and the result of the one-dimensional Fourier transform
- second frequency analysis unit 4 4 shown in step C 5 Operation of second frequency analysis unit 4 4 shown in step C 5, feature point detection shown in steps C 6 and C 7 Operation of Z ridge restoration unit 4 6, and calculation result output device 5 in step C 8
- the operations of 0 are the operations of the second frequency analysis unit 44, the feature point detection / ridge restoration unit 46, and the operation result output device 50 in the embodiment of FIG. 1 (step A5 in FIG. 4). , Steps A 6 and A 7 and steps A 8). That is, the computer 40 B further performs a one-dimensional Fourier transform on the transmitted one-dimensional Fourier transform result by the second frequency analysis unit 44, and obtains a two-dimensional Fourier transform result.
- the result of one-dimensional frequency analysis performed by dividing an input image (original image) into pixel columns is sequentially transmitted to the computer 40 B.
- the second frequency analysis unit 44 can be operated, and the processing can be started before the entire image is input.
- the image input device 30 B has a function of dividing the pixel row and a function of performing one-dimensional frequency analysis of each divided pixel row, and the first frequency from the computer 40 B side. Except for the point that the analysis unit is omitted, the overall configuration is substantially the same as that of the first embodiment of the present invention described with reference to FIGS.
- a microprocessor dedicated to signal processing such as DSP is mounted on the finger / palm-print image input device 30 B, and the first frequency analysis is performed by the finger / palm-print image input device 30 B.
- the processing of the computer 40 B can be reduced.
- the finger / palm-print image input device 30B is installed in a transceiver or other device that performs audio processing
- a new component can be created by using the processor dedicated to signal processing for audio processing that is originally installed in the transceiver. It is possible to reduce the processing of the computer 40 B without adding.
- only the second frequency analysis unit 44, or the first frequency analysis unit 42B and the second frequency analysis unit 44 can be configured by independent signal processing microprocessors. Is possible.
- a fingerprint or palm print image is assumed as an original image to be processed.
- Frequency analysis can also be performed.
- the Fourier transform has been described as an example of the frequency analysis method, other frequency analysis methods that can decompose the processing vertically and horizontally, such as Walsh-Hadamard transform, may be used.
- frequency analysis is performed on the pixels for one column to obtain one-dimensional frequency analysis results for each column, and these one-dimensional frequency analysis results are used together to obtain the final two-dimensional frequency analysis results. It is also possible to use other frequency analysis methods capable of obtaining
- an image processing system 20 C according to the fourth embodiment of the present invention will be briefly described with reference to FIG.
- the image processing system 20 C includes the first embodiment (FIG. 1), the second embodiment (FIG. 12), or the third embodiment (FIG.
- this is an image processing system including a computer 70 and an image input device 30 that operate according to the program.
- the program is read into the computer 70 and controls the operation of the CPU of the computer 70.
- the computer 70 is under the control of the image processing program and is the same as the processing by the computers 40, 4OA, 40B in the first embodiment, the second embodiment, or the third embodiment described first. Execute the process.
- the image processing program is as shown in FIG.
- the image processing program will be described with reference to FIG. This corresponds to the processing power S from step a1 to step a4, the function of sequentially fetching the original image sequentially and outputting the partially fetched image sequentially.
- the processing from step a5 to step a7 corresponds to the function of dividing the partially captured image into pixel columns and sequentially outputting the pixel columns.
- the processing from step a8 to a12 corresponds to the function of performing one-dimensional frequency analysis on each pixel column and sequentially obtaining the one-dimensional frequency analysis results.
- the processing from step a1 to step a12 corresponds to the first frequency angle analysis function that divides a partially captured image into pixel rows and performs one-dimensional frequency analysis on each pixel row. To do.
- the processing from step a 14 to step a 20 corresponds to a function that collects as many 1-dimensional frequency analysis results as necessary to analyze a small region and obtains a two-dimensional frequency analysis result of the small region. To do.
- the processing from step a 14 onwards collects the one-dimensional frequency analysis results obtained in the first frequency analysis necessary for the analysis of the small region, and obtains the two-dimensional frequency analysis results of the small region. This corresponds to the second frequency analysis function to be obtained.
- the first frequency analysis function that performs one-dimensional frequency analysis of each pixel row
- the force program is arranged on the side of the force device 30. It is technically easy to divide the force program and incorporate it into the computer 70 or the image input device 30.
- the image processing system according to the present invention is used for frequency analysis of finger / palm print images, it can be applied to applications such as a finger / palm print authentication device for identity verification and a program for realizing the finger / palm print authentication device on a computer. Also, it can be applied to applications such as a program for realizing a finger / palm print database registration device or a finger / palm print database registration device on a computer.
Abstract
Description
Claims
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US11/632,930 US7831071B2 (en) | 2004-07-22 | 2005-07-22 | Image processing system |
EP05767358A EP1788524B1 (en) | 2004-07-22 | 2005-07-22 | Image processing system |
DE602005013401T DE602005013401D1 (de) | 2004-07-22 | 2005-07-22 | Bildverarbeitungssystem |
KR1020077003967A KR100887029B1 (ko) | 2004-07-22 | 2005-07-22 | 화상 처리 시스템 |
JP2006527863A JP4569781B2 (ja) | 2004-07-22 | 2005-07-22 | 画像処理システム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004214919 | 2004-07-22 | ||
JP2004-214919 | 2004-07-22 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2006009319A1 true WO2006009319A1 (ja) | 2006-01-26 |
Family
ID=35785408
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2005/013890 WO2006009319A1 (ja) | 2004-07-22 | 2005-07-22 | 画像処理システム |
Country Status (7)
Country | Link |
---|---|
US (1) | US7831071B2 (ja) |
EP (2) | EP1788524B1 (ja) |
JP (1) | JP4569781B2 (ja) |
KR (1) | KR100887029B1 (ja) |
CN (1) | CN100573585C (ja) |
DE (2) | DE602005013401D1 (ja) |
WO (1) | WO2006009319A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009008446A (ja) * | 2007-06-26 | 2009-01-15 | Konica Minolta Holdings Inc | 情報処理システム、プログラム、および情報処理方法 |
JP2010123020A (ja) * | 2008-11-21 | 2010-06-03 | Rakuten Inc | 目標特徴線特定装置、情報出力装置、目標特徴線特定方法、及び目標特徴線特定処理プログラム |
WO2017022548A1 (ja) * | 2015-07-31 | 2017-02-09 | 株式会社ディー・ディー・エス | 情報処理プログラム及び情報処理装置 |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006009319A1 (ja) * | 2004-07-22 | 2006-01-26 | Nec Corporation | 画像処理システム |
US8306288B2 (en) | 2009-08-19 | 2012-11-06 | Harris Corporation | Automatic identification of fingerprint inpainting target areas |
JP5857256B2 (ja) * | 2010-11-30 | 2016-02-10 | パナソニックIpマネジメント株式会社 | オブジェクト検出装置、画像分割装置、集積回路 |
US10559076B2 (en) * | 2015-03-11 | 2020-02-11 | Fuji Corporation | Component type automatic distinguishing method, component type automatic distinguishing system, image processing component data creation method, and image processing component data creation system |
JP6657593B2 (ja) * | 2015-05-08 | 2020-03-04 | 富士通株式会社 | 生体撮影装置、生体撮影方法及び生体撮影プログラム |
USD776664S1 (en) * | 2015-05-20 | 2017-01-17 | Chaya Coleena Hendrick | Smart card |
KR102452945B1 (ko) * | 2015-08-27 | 2022-10-11 | 삼성전자주식회사 | 푸리에 변환을 수행하는 방법 및 장치 |
KR20180051441A (ko) * | 2016-09-27 | 2018-05-16 | 선전 구딕스 테크놀로지 컴퍼니, 리미티드 | 지문 인식 시스템 |
FR3084945B1 (fr) * | 2018-08-10 | 2020-08-07 | Idemia Identity & Security France | Procede d'analyse d'une empreinte |
EP3980933A4 (en) * | 2019-06-04 | 2023-08-02 | IDEMIA Identity & Security USA LLC | DIGITAL IDENTIFIER FOR A DOCUMENT |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5922173A (ja) * | 1982-07-28 | 1984-02-04 | Nippon Telegr & Teleph Corp <Ntt> | 指紋画像コ−ド化方法 |
JPH07271763A (ja) * | 1994-03-30 | 1995-10-20 | Olympus Optical Co Ltd | データ解析装置 |
JPH08131419A (ja) * | 1994-11-14 | 1996-05-28 | Hitachi Medical Corp | 磁気共鳴イメージング装置 |
JPH09214967A (ja) * | 1996-01-30 | 1997-08-15 | Fuji Photo Film Co Ltd | 画像データ圧縮処理方法 |
JP2001136526A (ja) * | 1999-11-05 | 2001-05-18 | Canon Inc | 画像処理方法及びその装置及び記憶媒体 |
JP2003295990A (ja) * | 2002-04-05 | 2003-10-17 | Sharp Corp | 機器の操作装置及び方法 |
Family Cites Families (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3779492A (en) * | 1971-10-18 | 1973-12-18 | Grumman Aerospace Corp | Automatic target recognition system |
US3836712A (en) * | 1972-12-29 | 1974-09-17 | S Kowel | Direct electronic fourier transforms of optical images |
JPS6059473A (ja) * | 1983-09-09 | 1985-04-05 | Yokogawa Hokushin Electric Corp | 投影波形生成回路 |
JP3057590B2 (ja) * | 1992-08-06 | 2000-06-26 | 中央発條株式会社 | 個人識別装置 |
JPH09147113A (ja) * | 1995-11-27 | 1997-06-06 | Nippon Telegr & Teleph Corp <Ntt> | 指紋照合方法及び装置 |
US5859920A (en) * | 1995-11-30 | 1999-01-12 | Eastman Kodak Company | Method for embedding digital information in an image |
JP2739856B2 (ja) | 1995-12-18 | 1998-04-15 | 日本電気株式会社 | 指掌紋画像処理装置 |
FR2749955B1 (fr) | 1996-06-14 | 1998-09-11 | Thomson Csf | Systeme de lecture d'empreintes digitales |
US5953442A (en) * | 1997-07-24 | 1999-09-14 | Litton Systems, Inc. | Fingerprint classification via spatial frequency components |
JP3036497B2 (ja) * | 1997-12-01 | 2000-04-24 | 日本電気株式会社 | 画像データの拡大率・縮小率検出装置、及び検出方法 |
JP3700381B2 (ja) | 1998-03-18 | 2005-09-28 | コニカミノルタビジネステクノロジーズ株式会社 | 画像処理装置 |
US7227884B2 (en) * | 2000-02-28 | 2007-06-05 | Aeroastro, Inc. | Spread-spectrum receiver with progressive fourier transform |
JP2004505340A (ja) | 2000-07-19 | 2004-02-19 | キム,ヤン ワン | カードレス安全信用取引処理のためのシステムおよび方法 |
JP2002099912A (ja) | 2000-09-21 | 2002-04-05 | Nippon Telegr & Teleph Corp <Ntt> | 画像処理方法,画像処理装置および画像処理プログラム記録媒体 |
JP4193163B2 (ja) | 2001-03-26 | 2008-12-10 | 日本電気株式会社 | 指掌紋画像処理装置及び方法 |
JP4134522B2 (ja) * | 2001-03-26 | 2008-08-20 | 日本電気株式会社 | 指掌紋画像処理装置及び方法 |
FR2831302A1 (fr) * | 2001-10-19 | 2003-04-25 | St Microelectronics Sa | Codage d'informations concentriques |
US7013030B2 (en) * | 2002-02-14 | 2006-03-14 | Wong Jacob Y | Personal choice biometric signature |
JP2003331337A (ja) * | 2002-05-16 | 2003-11-21 | Ricoh Co Ltd | カラー画像判定装置 |
JP3811105B2 (ja) | 2002-07-29 | 2006-08-16 | 株式会社駅前探険倶楽部 | 経路探索方法及び経路探索プログラム |
WO2005086091A1 (ja) * | 2004-03-04 | 2005-09-15 | Nec Corporation | 指紋掌紋画像処理システムおよび指紋掌紋画像処理方法 |
WO2006009319A1 (ja) * | 2004-07-22 | 2006-01-26 | Nec Corporation | 画像処理システム |
-
2005
- 2005-07-22 WO PCT/JP2005/013890 patent/WO2006009319A1/ja active Application Filing
- 2005-07-22 EP EP05767358A patent/EP1788524B1/en not_active Expired - Fee Related
- 2005-07-22 CN CNB2005800296368A patent/CN100573585C/zh not_active Expired - Fee Related
- 2005-07-22 US US11/632,930 patent/US7831071B2/en not_active Expired - Fee Related
- 2005-07-22 DE DE602005013401T patent/DE602005013401D1/de active Active
- 2005-07-22 EP EP08157168A patent/EP1967988B1/en not_active Expired - Fee Related
- 2005-07-22 DE DE602005020530T patent/DE602005020530D1/de active Active
- 2005-07-22 JP JP2006527863A patent/JP4569781B2/ja not_active Expired - Fee Related
- 2005-07-22 KR KR1020077003967A patent/KR100887029B1/ko not_active IP Right Cessation
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5922173A (ja) * | 1982-07-28 | 1984-02-04 | Nippon Telegr & Teleph Corp <Ntt> | 指紋画像コ−ド化方法 |
JPH07271763A (ja) * | 1994-03-30 | 1995-10-20 | Olympus Optical Co Ltd | データ解析装置 |
JPH08131419A (ja) * | 1994-11-14 | 1996-05-28 | Hitachi Medical Corp | 磁気共鳴イメージング装置 |
JPH09214967A (ja) * | 1996-01-30 | 1997-08-15 | Fuji Photo Film Co Ltd | 画像データ圧縮処理方法 |
JP2001136526A (ja) * | 1999-11-05 | 2001-05-18 | Canon Inc | 画像処理方法及びその装置及び記憶媒体 |
JP2003295990A (ja) * | 2002-04-05 | 2003-10-17 | Sharp Corp | 機器の操作装置及び方法 |
Non-Patent Citations (1)
Title |
---|
See also references of EP1788524A4 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009008446A (ja) * | 2007-06-26 | 2009-01-15 | Konica Minolta Holdings Inc | 情報処理システム、プログラム、および情報処理方法 |
JP2010123020A (ja) * | 2008-11-21 | 2010-06-03 | Rakuten Inc | 目標特徴線特定装置、情報出力装置、目標特徴線特定方法、及び目標特徴線特定処理プログラム |
WO2017022548A1 (ja) * | 2015-07-31 | 2017-02-09 | 株式会社ディー・ディー・エス | 情報処理プログラム及び情報処理装置 |
KR20180035824A (ko) * | 2015-07-31 | 2018-04-06 | 가부시끼가이샤 디디에스 | 정보 처리 프로그램 및 정보 처리 장치 |
JPWO2017022548A1 (ja) * | 2015-07-31 | 2018-05-17 | 株式会社ディー・ディー・エス | 情報処理プログラム及び情報処理装置 |
US10430629B2 (en) | 2015-07-31 | 2019-10-01 | Kabushiki Kaisha Dds | Non-transitory computer-readable medium storing information processing program and information processing device |
KR102457004B1 (ko) | 2015-07-31 | 2022-10-21 | 가부시끼가이샤 디디에스 | 정보 처리 프로그램 및 정보 처리 장치 |
Also Published As
Publication number | Publication date |
---|---|
EP1967988B1 (en) | 2010-04-07 |
DE602005020530D1 (de) | 2010-05-20 |
JP4569781B2 (ja) | 2010-10-27 |
CN101010697A (zh) | 2007-08-01 |
JPWO2006009319A1 (ja) | 2008-05-01 |
CN100573585C (zh) | 2009-12-23 |
DE602005013401D1 (de) | 2009-04-30 |
EP1788524B1 (en) | 2009-03-18 |
KR20070035097A (ko) | 2007-03-29 |
EP1788524A1 (en) | 2007-05-23 |
EP1967988A1 (en) | 2008-09-10 |
KR100887029B1 (ko) | 2009-03-04 |
US20070297655A1 (en) | 2007-12-27 |
US7831071B2 (en) | 2010-11-09 |
EP1788524A4 (en) | 2008-03-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP4569781B2 (ja) | 画像処理システム | |
JP5206218B2 (ja) | 指紋画像取得装置、指紋認証装置、指紋画像取得方法及び指紋認証方法 | |
JP4029412B2 (ja) | 指紋掌紋画像処理システムおよび指紋掌紋画像処理方法 | |
JP5184682B2 (ja) | 指紋プリビュー品質ならびに細分化 | |
JP5449329B2 (ja) | 強化された画像整合のための方法およびシステム | |
CN106663198A (zh) | 用于超声波指纹的基于图像的活性检测 | |
EP2104059B1 (en) | Apparatus and method for processing image | |
US9607204B2 (en) | Biometric information processing apparatus, method, and recording medium | |
Birajadar et al. | Towards smartphone-based touchless fingerprint recognition | |
JP2021532453A (ja) | フィードフォワード畳み込みニューラルネットワークを使用した高速且つ堅牢な皮膚紋理の印のマニューシャの抽出 | |
US9679218B2 (en) | Method and apparatus for image matching | |
WO2019228471A1 (zh) | 指纹识别方法、设备及计算机可读存储介质 | |
JP4411152B2 (ja) | 指紋読取方法、指紋読取システム及びプログラム | |
JP6229352B2 (ja) | 画像処理装置、画像処理方法およびプログラム | |
US8655084B2 (en) | Hand-based gender classification | |
KR102558736B1 (ko) | 지문 인식 방법 및 장치 | |
Oblak et al. | Fingermark quality assessment framework with classic and deep learning ensemble models | |
Masmoudi et al. | Implementation of a fingerprint recognition system using LBP descriptor | |
JP3164835B2 (ja) | 人物画像認識における前処理方法および後処理方法 | |
US8358803B2 (en) | Navigation using fourier phase technique | |
JP2006330873A (ja) | 指紋照合装置、方法およびプログラム | |
Prasad et al. | Proposed Biometric Iris Recognition System Implementation Using Matlab | |
CN113688691A (zh) | 一种人脸识别系统及人脸识别方法 | |
HARIA | Touchless Fingerprint Recognition System | |
Moua et al. | Edge detection and feature extraction in automated fingerprint identification systems |
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 KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM 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): 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 LV 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 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
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: 2006527863 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 11632930 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020077003967 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2005767358 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200580029636.8 Country of ref document: CN |
|
WWP | Wipo information: published in national office |
Ref document number: 1020077003967 Country of ref document: KR |
|
WWP | Wipo information: published in national office |
Ref document number: 2005767358 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 11632930 Country of ref document: US |