US RE37091 E1 Abstract A motion compensated prediction interframe coding system which first measures characteristics regarding the fineness of a pattern or texture of a picture, which is represented by an input television signal and is divided into a group of continuous blocks each quantized by using a quantization step size, and changes the quantization step size into a smaller one if quantizes a block having a picture pattern or texture finer than patterns of the other blocks. Thereby, an amount of generated codes can be limited but a motion compensated prediction frame coding operation can be performed without degrading the fineness of the texture of the original input picture.
Claims(33) 1. A motion compensated prediction interframe coding system comprising:
an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal;
a block dividing means for dividing a predetermined area of the television signal digitalized by said analog-to-digital conversion means into blocks each having a predetermined size;
a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not the motion compensation is performed by using by the motion vector;
a motion compensation means for performing a motion compensation of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation is determined by said motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation is effected;
a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof;
an orthogonal transform means for performing an orthogonal transform of the prediction error of each pixel of a coding block of the television picture so as to obtain orthogonal transform coefficients;
a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes;
a variance calculating means for calculating a variance of the gray levels of the pixels of each block of input television signal;
a second quantization step-size computing means for classifies the blocks into a predetermined number of classes according to the variance and computing a second step size for the quantization of each block from the first step size;
a quantization means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of orthogonal transform;
an orthogonal transform coefficient coding means for performing the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform;
a quantized prediction error calculating means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain quantized prediction errors;
a reproduced picture calculating means for calculating a reproduced picture from the quantized prediction errors and the predicted gray levels of the pixels of the blocks;
a memory portion for storing the reproduced picture; and
a motion vector coding means for performing the coding of the motion vectors.
2. A motion compensated prediction interframe coding system as set forth in claim
1 wherein the predetermined area is one frame.3. A motion compensated prediction interframe coding system as set forth in claim
1 wherein the predetermined area is one field.4. A motion compensated prediction interframe coding system comprising:
an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal;
a block dividing means for dividing a predetermined area of the television signal digitalized by said analog-to-digital conversion means into blocks each having a predetermined size;
a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not the motion compensation is performed by using by the motion vector;
a motion compensation means for performing a motion compensation of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation is determined by said motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation is effected;
a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof;
a coding method selection means for selecting a method of coding of a block from an interframe coding and intraframe coding methods;
a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by said coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors;
an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by said first switch means;
a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes;
a variance calculating means for calculating a variance of the gray levels of the pixels of each block of input television signal;
a second quantization step-size computing means for classifying the blocks into a predetermined number of classes according to the variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intra-frame coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method;
a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transforms by using the second step size to obtain quantized coefficients of an orthogonal transform;
an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform;
an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain inversely-quantized values;
a second switch means for selecting a numerical value of zero as a gray level of a pixels of a reproduced picture in case where said coding method selection means selects the intraframe coding method and in contrast, selects the predicted value as the gray level of a pixel of the reproduced picture in case where the coding method selection means selects the interframe coding method;
a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks;
a memory portion for storing the reproduced picture; and
a motion vector coding means for performing the coding of the motion vectors.
5. A motion compensated prediction interframe coding system as set forth in claim
4 wherein the predetermined area, is one frame.6. A motion compensated prediction interframe coding system as set forth in claim
4 wherein the predetermined area is one field.7. A motion compensated prediction interframe coding system comprising:
an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal;
a block dividing means for dividing a predetermined area of the television signal digitalized by said analog-to-digital conversion means into blocks each having a predetermined size;
a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not a motion compensation prediction is performed by using by the motion vector;
a motion compensation predicting means for performing a motion compensation prediction of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation prediction is determined by said motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation prediction is effected;
a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof;
a coding method selection means for selecting a method of coding of a block from an interframe coding and intraframe coding methods;
a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by said coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors;
an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by said first switch means;
a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes;
an average and variance calculating means for calculating an averaged value and a variance of the gray levels of the pixels of each block of input television signal;
a second quantization step-size computing means for classifying the blocks into a predetermined number of classes according to the averaged value and variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intra-frame coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method;
a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of an orthogonal transform;
an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform;
an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain inversely-quantized values;
a second switch means for selecting a numerical value of zero or the predicted gray level obtained by the motion compensation prediction as a gray level of a pixels of a reproduced picture according to the selection made by said coding method selection means;
a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks;
a memory portion for storing the reproduced picture; and
a motion vector coding means for performing the coding of the motion vectors.
8. A motion compensated prediction interframe coding system as set forth in claim
7 wherein the predetermined area is one frame.9. A motion compensation prediction interframe coding system as set forth in claim
7 wherein the predetermined area is one field.10. A motion compensated prediction interframe coding system comprising:
a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not a motion compensation prediction is performed by using by the motion vector;
a motion compensation predicting means for performing a motion compensation prediction of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation prediction is determined by said motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation prediction is effected;
a coding method selection means for selecting a method of coding of a block from an intraframe coding and interframe coding methods;
a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by said coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors;
an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by said first switch means;
an average and variance calculating means for calculating an averaged value and a variance of the gray levels of the pixels of each block of input television signal;
a threshold value modifying means for modifying predetermined threshold values according to the first step size to classify blocks, which are represented by luminance signals and are to be coded by effecting the intraframe coding method, into a predetermined number of classes;
a second quantization step-size computing means for classifying the blocks, which are represented by luminance signals and are to be coded by effecting the intraframe coding method, into the predetermined number of classes according to the averaged value and variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intraframe coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method;
a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of an orthogonal transform;
an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform;
an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain inversely-quantized values;
a second switch means for selecting a numerical value of zero or the predicted gray level obtained by the motion compensation prediction as a gray level of a pixels of a reproduced picture according to the selection made by said coding method selection means;
a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks;
a memory portion for storing the reproduced picture; and
a motion vector coding means for performing the coding of the motion vectors.
11. A motion compensated prediction interframe coding system as set forth in claim
10 wherein the predetermined area is one frame.12. A motion compensated prediction interframe coding system as set forth in claim
10 wherein the predetermined area is one field.13. A motion compensated prediction interframe coding system as set forth in claim
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 wherein the variance is calculated from the following equation: where M represents the number of pixels on a row of a block; N the number of rows of the block; p(i,j) a gray level of a pixel at an address (i,j) in the block; and P the averaged value of the gray levels of the pixels of the block obtained by
14. A motion compensated prediction interframe coding system as set forth in claim
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 wherein the variance is calculated from the following equation: 15. A motion compensated prediction interframe coding system as set forth in claim
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 wherein the predetermined number of the classes is four.16. A motion compensated prediction interframe coding system as set forth in claim
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12 wherein the second step size is determined to be small for a block having a small variance.17. A motion compensated prediction interframe coding system as set forth in claim
10, 11 or 12 wherein the threshold values are modified by using the following equation:where a subscript i indicates an integer from 1 to 4; and “max” represents a maximum value of the first step size.
18. A moving-
image signal encoding apparatus comprising: a memory portion for storing codes;
means for computing a first step size for quantization from an amount of generated codes or from occupied capacity of the memory portion;
means for modifying the first quantization step size to obtain a second quantization step size; and
means for encoding a signal quantized in accordance with the second quantization step size.
19. A moving-
image signal encoding apparatus as set forth in claim 18, wherein the second quantization step size is computed from the first quantization step size and a local image content. 20. A moving-
image signal encoding apparatus as set forth in claim 19, wherein the local image content is judged from a value of a variance of an input television signal. 21. The moving-
image signal encoding apparatus as set forth in claim 19, wherein the local image content is judged from both of a value of a variance and a value of average which are calculated from an input television signal. 22. A moving-
image signal encoding apparatus comprising: a memory portion for storing codes;
means for computing a first step size for the quantization of a signal from an amount of codes remained in the memory portion;
means for modifying the first quantization step size to obtain a second quantization step size;
means for quantizing the signal in accordance with the second quantization step size; and
means for encoding the quantized signal.
23. The moving-
image signal encoding apparatus as set forth in claim 22, wherein the means for modifying the first quantization size computes the second quantization step size from both of the first quantization step size and a value of a local image characteristic measure which represents local characteristics of an input image represented by an input television signal. 24. The moving-
image signal encoding apparatus as set forth in claim 23, wherein the local image characteristic measure is a variance of the input television signal. 25. The moving-
image signal encoding apparatus as set forth in claim 23, wherein the local image characteristic measure is an average of the input television signal. 26. A method comprising the steps of:
computing a first step size for quantization from an amount of generated codes or from occupied capacity of a memory portion;
modifying the first quantization step size to obtain a second quantization step size; and
encoding a signal quantized in accordance with the second quantization step size.
27. The method as set forth in claim
26, wherein the step of modifying the first quantization step size comprises the sub-step of computing the second quantization step size from the first quantization step size and a local image content. 28. The method as set forth in claim
27, wherein the step of modifying the first quantization step size further comprises the sub-step of nudging the local image content from a value of a variance of an input television signal. 29. The method as set forth in claim
27, wherein the step of modifying the first quantization step size further comprises the sub-step of judging the local image content from both of a value of a variance and a value of average which are calculated from an input television signal. 30. A method comprising the steps of:
computing a first step size for the quantization of a signal from an amount of codes remained in a memory portion;
modifying the first quantization step size to obtain a second quantization step size;
quantizing the signal in accordance with the second quantization step size; and
encoding the quantized signal.
31. The method as set forth in claim
30, wherein the step of modifying the first quantization step size comprises the sub-step of computing the second quantization step size from both of the first quantization step size and a value of a local image characteristic measure which represents local characteristics of an input image represented by an input television signal. 32. The method as set forth in claim
31, wherein the local image characteristic measure is a variance of the input television signal. 33. The method as set forth in claim
31, wherein the local image characteristic measure is an average of the input television signal. Description 1. Field of the Invention This invention generally relates to a television system and more particularly to a motion compensated prediction interframe system for performing motion compensated prediction interframe coding of television signals. 2. Description of the Related Art Recently, with advance in techniques of moving picture coding, a motion compensated prediction interframe coding system has been developed as an efficient coding system for effecting an efficient coding of a color moving picture for use in a visual telephone (or video telephone), a video conference system, CD ROM, a digital video tape recorder (VTR) and so on. An example of a motion compensated prediction interframe coding system is described in T. Fukinuki: Multidimensional Signal Processing of TV Picture, Nikkan Kogyo Shinbun Company, Chapter 7 “Efficient Coding”, pp. 213-291 (1988. 11. 15). When a large quantity of prediction error signals is generated, the motion compensated prediction interframe coding system limits an amount of codes of the generated prediction error signals by employing a large step size for quantization (hereunder sometimes referred to as a quantization step size) in such a way that at the coding of a picture can be achieved a constant frame rate. An example of a conventional method of determining a step size for quantization is disclosed in CCITT (Comite Consultatif Internationale Telegraphique et Telephinque) SGXV Document #525: Description of Ref. Model 8 (RM 8), Working Party XV/4 Specialist Group on Coding for Visual Telephony (1989. 6. 9). Hereinafter, a prior art motion compensated prediction interframe coding system will be describe by referring to FIG. In this figure, reference numeral Hereinafter, an operation of the prior art motion compensated prediction interframe coding system having the above described arrangement will be described. First, television signals are converted into digital television signals by an analog-to-digital (A/D) conversion circuit (not shown). Then, pixels represented by the digital television signals are divided into blocks each of which is a rectangular array composed of M×N pixels arranged in M columns and N rows. Further, the digital television signals are input to the system from the input terminal The motion compensation prediction portion The coding method selection portion (1) Picture quality can be improved after a scene is changed because the intraframe coding method is employed when the scene is changed. (2) Picture quality can be also improved when a moving body largely moves because a background which has hid itself behind the moving body emerges from behind it and in such a case, the intraframe coding is employed. Incidentally, in case of a coding method employed in storage type media such as CD ROM, it is necessary to insert a frame of which all blocks are coded by performing the intraframe coding method (hereunder referred to as a refreshment frame) between frames every constant frame period in order to realize an editing function and a backward editing function of a reproduced picture. Further, the insertion of the refreshment frame can be achieved by providing the motion compensation prediction interframe coding system with the intraframe coding function. Then, the intra-loop filtering portion Thereafter, in case where the coding method selection signal Further, the orthogonal transform portion Then, the quantization portion The quantization step-size computing portion The method of calculating the quantization step size In this prior art system, a picture represented by the input television signal is composed of 352 columns and 288 rows (i.e., 352 pixels in each row and 288 lines in the vertical direction) and is divided into blocks each of which is referred to as a Macro Block and is composed of 16 columns and 16 rows (i.e., 16 pixels in each row and 16 lines in the vertical direction), as illustrated in FIG.
where (a) INT[*] is defined as a function of rounding a part after the decimal point of an argument * down (e.g., INT[1.5]=1, INT[1.3]=1, INT[1.6]=1), (b) Bcont represents the amount of the code remained in the code memory portion
(e.g., q=1 if V=64K bits/second). As is apparent from the equation (1), when the amount of the remained code Bcont increases, the quantization step size Qb also increases so as to limit an amount of generated codes and achieve the coding of video signals at a constant frame rate. For instance, if the amount of the remained code Bcont=700 bits, the quantization step size Qb=8. Further, if the amount of the remained code Bcont=6100 bits, the quantization step size Qb=62. Incidentally, the quantization is performed by using a predetermined quantization step size Qb with respect to a first Macro Block to a (n−1)th Macro Block. For example, the Qb is set to be 32 in case where V=64K bits/sec. and q=1. In this prior art system, the value of n representing an interval of the calculations of the quantization step size is set to be 12. The inverse orthogonal transform portion Further, the switching portion Then, the reproduced picture calculating portion Furthermore, the motion vector coding portion However, in this prior art system constructed as above described, the quantization step size Qb is fixed during an interval of blocks (n blocks (i.e., n Marco Blocks) in case of this prior art system) at which the calculation of the quantization step size is calculated. That is, prediction errors are quantized by using the same quantization step size Qb during the continuous n blocks independently of characteristics of the input television signals. More particularly, during the interval at which the calculation of the quantization is calculated, of the continuous blocks, a block having a fine pattern or texture and another block having a coarse pattern are quantized by using the same quantization step size Qb. Thus, the prior art system has a first drawback that the picture quality of the block having a fine patter is degraded, namely, what is called a “block distortion”, which is a phenomenon that the block loses the fineness of the pattern thereof and comes to have a flat pattern, occurs due to the quantization of a prediction error generated from the block having a fine pattern and is visually perceived as degradation in picture quality. Further, similarly, the quantized orthogonal transform coefficients are quantized by using the same quantization step size Qb during the continuous n blocks independently of characteristics of the input television signals. That is during the interval at which the calculation of the quantization is calculated, of the continuous blocks, a block having a fine pattern, which should be coded by using the interframe coding method, and another block having a coarse pattern are quantized by using the same quantization step size Qb. Thus, the prior art system has another similar drawback (hereunder referred to as a second drawback) that the picture quality of the block, which should be coded by the interframe coding method, having a fine patter is degraded, namely, the “block distortion” occurs due to the quantization of orthogonal transform coefficients generated from the block having a fine pattern and is visually perceived as degradation in picture quality. On the other hand, in order to reduce the amount of the codes generated by coding the quantization step size, the number of the continuous blocks quantized by using the same quantization step size needs to be more than a predetermined number, which is n in case of the above described prior art system. Thus, practically, it is not possible that the quantization step size is calculated or changed every block. The present invention is created to obviate the above described drawbacks of the prior art system. It is therefore a first object of the present invention to eliminate the first drawback of the prior art system, namely, to provide a motion compensated prediction interframe coding system which can improve the picture quality by calculating a second quantization step size for each block from a reference (i.e., a first quantization step size), which is calculated from an amount of generated codes as proportional to the visual fineness of each of the continuous blocks to be quantized by using the first quantization step size, and further quantizing the prediction error signals by using the second quantization step size, namely, by employing a second quantization step size, which is smaller than the first quantization step size, for the quantization of a block having a fine pattern to limit the amount of the generated codes and maintain the fineness of the block and to thereby improve the picture quality of the block having a fine pattern, as well as the picture quality of an entire picture. To achieve the foregoing first object and in accordance with a first aspect of the present invention, there is provided a motion compensated prediction interframe coding system which comprises an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal, a block dividing means for dividing a predetermined area of the television signal digitalized by the analog-to-digital conversion means into blocks each having a predetermined size, a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each whether or not the motion compensation is performed by using by the motion vector, a motion compensation means for performing a motion compensation of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation is determined by the motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation is effected, a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof, an orthogonal transform means for performing an orhtogonal transform of the prediction error of each pixel of a coding block of the television picture so as to obtain orthogonal transform coefficients; a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes, a variance calculating means for calculating a variance of the gray levels of the pixels of each block of input television signal, a second quantization step-size computing means for classifies the blocks into a predetermined number of classes according to the variance and computing a second step size for the quantization of each block from the first step size, a quantization means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of orthogonal transform, an orthogonal transform coefficient coding means for performing the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform, a quantized prediction error calculating means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain quantized prediction errors, a reproduced picture calculating means for calculating a reproduced picture from the quantized prediction errors and the predicted gray levels of the pixels of the blocks, a memory portion for storing the reproduced picture and a motion vector coding means for performing the coding of the motion vectors. Incidentally, the fineness of the pattern of a block of the input television signal is considered as can be represented by a variance σ Thus, the motion compensated prediction interframe coding system according to the first aspect of the present invention can limit the amount of the generated codes and improve the picture quality of a block having a fine pattern by measuring the fineness of patterns of blocks of the input television signal in terms of the variance σ Further, it is a second object of the present invention to eliminate the second drawback of the prior art system, namely, to provide a motion compensated prediction interframe coding system which can improve the picture quality by calculating a second quantization step size for blocks to be coded by using the intraframe coding method from a reference (i.e., a first quantization step size), which is calculated from an amount of generated codes as proportional to the visual fineness of the blocks to be coded by using the intraframe coding method and further quantizing the orthogonal transform coefficients by using the second quantization step size, namely, by employing a second quantization step size, which is smaller than the first quantization step size, for the quantization of a block having a fine pattern to limit the amount of the generated codes and maintain the fineness of the block and to thereby improve the picture of the block having a fine pattern, as well as the picture quality of an entire picture. To achieve the foregoing second object and in accordance with a second aspect of the present invention, there is provided a motion compensated prediction interframe coding system which comprises an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal, a block dividing means for dividing a predetermined area of the television signal digitalized by the analog-to-digital conversion means into blocks each having a predetermined size, a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not the motion compensation is performed by using by the motion vector, a motion compensation means for performing a motion compensation of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation is determined by the motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation is effected, a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof, a coding method selection means for selecting a method of coding of a block from an interframe coding and interframe coding methods, a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by the coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors, an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by the first switch means, a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes, a variance calculating means for calculating a variance of the gray levels of the pixels of each block of input television signal, a second quantization step-size computing means for classifying the blocks into a predetermined number of classes according to the variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intra-frame coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method, a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of an orthogonal transform, an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform, an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain inversely-quantized values, a second switch means for selecting a numerical value of zero as a gray level of a pixels of a reproduced picture in case where the coding method selection means selects the intraframe coding method and in contrast, selects the predicted value as the gray level of a pixel of the reproduced picture in case where the coding method selection means selects the interframe coding method, a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks, a memory portion for storing the reproduced picture and a motion vector coding means for performing the coding of the motion vectors. As above described, the fineness of the pattern of a block of the input television signal is considered as can be represented by a variance σ Therefore, the motion compensated prediction interframe coding system according to the second aspect of the present invention can limit the amount of the generated codes and improve the picture quality of a block, which should be coded by using the intraframe coding method and has a fine pattern, by measuring the fineness of patterns of blocks of the input television signal in terms of the variance σ Moreover, it is a third object of the present invention to eliminate the second drawback of the prior art system, namely, to provide a motion compensated prediction interframe coding system which can maintain the fineness of an original picture and thereby improve the picture quality of the reproduced picture. To achieve the foregoing third object and in accordance with a third aspect of the present invention, there is provided a motion compensated prediction interframe coding system which comprises an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal, a block dividing means for dividing a predetermined area of the television signal digitalized by the analog-to-digital conversion means into blocks each having a predetermined size, a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not a motion compensation prediction is performed by using by the motion vector, a motion compensation predicting means for performing a motion compensation prediction of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation prediction is determined by the motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation prediction is effected, a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof, a coding method selection means for selecting a method of coding of a block from an interframe coding and interframe coding methods, a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by the coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors, an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by the first switch means, a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes, an average and variance calculating means for calculating an averaged value and a variance of the gray levels of the pixels of each block of input television signal, a second quantization step-size computing means for classifying the blocks into a predetermined number of classes according to the averaged value and variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intra-frame coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method, a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of an orthogonal transform, an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform, an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orthogonal transform to obtain inversely-quantized values, a second switch means for selecting a numerical value of zero or the predicted gray level obtained by the motion compensation prediction as a gray level of a pixels of a reproduced picture according to the selection made by the coding method selection means, a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks, a memory portion for storing the reproduced picture and a motion vector coding means for performing the coding of the motion vectors. As stated above, the fineness of the pattern of a block of the input television signal is considered as can be represented by a variance σ Further, in case where the averaged value of the the gray levels of pixels of a block is a certain value, this block may not visually be striking even if the variance of the gray levels of the pixels of this block is small and the fineness of the pattern of this block is high. Therefore, the motion compensated prediction interframe coding system according to the second aspect of the present invention can limit the amount of the generated codes and improve the picture quality of a block, which should be coded by using the intraframe coding method and has a fine pattern and of which the averaged value of gray levels of pixels is greater or equal to a predetermined value, by measuring the fineness of a pattern of each block of the input television signal in terms of the variance σ Furthermore, it is a fourth object of the present invention to eliminate the second drawback of the prior art system, namely, to provide a motion compensated prediction interframe coding system which can improve the picture quality by calculating a second quantization step size for each block (hereunder sometimes referred to as luminance signal block), which is to be coded by using the intraframe coding method and respectively corresponds to a part of a luminance signal and of which the averaged value of the gray levels of corresponding pixels is equal to or more than a predetermined value, from a reference (i.e., a first quantization step size), which is calculated from an amount of generated codes as proportional to the visual fineness of the blocks to be coded by using the intraframe coding method and further quantizing the orthogonal transform coefficients by using the second quantization step size, namely, by employing a second quantization step size, which is smaller than the first quantization step size, for the quantization of a luminance signal block, of which the averaged value of gray levels of corresponding pixels is equal to or more than a predetermined value and the fineness is high, to limit the amount of the generated codes, to maintain the fineness of the original picture and to thereby improve the picture quality of the blocks, and which can prevent discoloration of a coding block, which is due to the fact that an amount of codes corresponding to a color difference signal is extremely small, by reducing the amount of codes corresponding to the luminance signal and increase an amount of codes corresponding to the color difference signal by modifying a threshold value for evaluating the second quantization step size when the first quantization step size is equal to or more than a predetermined value, thereby improving the picture quality of the entire reproduced picture. To achieve the foregoing fourth object and in accordance with a fourth aspect of the present invention, there is provided a motion compensated prediction interframe coding system which comprises an analog-to-digital conversion means for performing an analog-to-digital conversion of a television signal, a block dividing means for dividing a predetermined area of the television signal digitalized by the analog-to-digital conversion means into blocks each having a predetermined size, a motion vector calculating means for calculating a motion vector, which represents a motion of a television picture represented by the television signal, of each block and for judging with respect to each block whether or not a motion compensation prediction is performed by using by the motion vector, a motion compensation predicting means for performing a motion compensation prediction of a reproduced picture signal of a previous frame with respect to each block of which the motion compensation prediction is determined by the motion vector calculating means to be effected and for calculating predicted gray levels of pixels of each block of which the motion compensation prediction is effected, a prediction error evaluating means for computing the difference between the gray level of each pixel of a coding block of the television picture and the predicted gray level thereof as a prediction error thereof, a coding method selection means for selecting a method of coding of a block from an interframe coding and interframe coding methods, a first switch means for selecting a signal, of which an orthogonal transform is effected, in accordance with the selection made by the coding method selecting means from the gray levels of pixels of blocks represented by the television signal and the prediction errors, an orthogonal transform means for performing an orthogonal transform of the prediction error or the gray levels selected by the first switch means; a first quantization step-size computing means for computing a first step size for the quantization from an amount of generated codes, an average and variance calculating means for calculating an averaged value and a variance of the gray levels of the pixels of each block of input television signal, a threshold value modifying means for modifying predetermined threshold values according to the first step size to classify blocks, which are repressented by luminance signals and are to be coded by effecting the intraframe coding method, into a predetermined number of classes, a second quantization step-size computing means for classifying the blocks, which are represented by luminance signals and are to be coded by effecting the intraframe coding method, into the predetermined number of classes according to the averaged value and variance and computing a second step size for the quantization of each block from the first step size in case of the blocks to be coded by using an intraframe coding method and for treating the first step sizes as the second step sizes in case of the blocks to be coded by using an interframe coding method, a quantized transform coefficient calculating means for quantizing coefficients of an orthogonal transform by using the second step size to obtain quantized coefficients of an orthogonal transform, an orthogonal transform coefficient coding means for performing information on the selection of the coding methods, the coding of the prediction error, the first step size and information on the classes and the quantized coefficients of the orthogonal transform, an inverse quantization means for effecting an inverse orthogonal transform of the quantized coefficients of the orhtogonal transform to obtain inversely-quantized values, a second switch means for selecting a numerical value of zero, or the predicted gray level obtained by the motion compensation prediction as a gray level of a pixels of a reproduced picture according to the selection made by the coding method selection means, a reproduced picture calculating means for calculating a reproduced picture from the inversely-quantized values and the numerical value of zero or from the predicted gray levels of the pixels of the blocks, a memory portion for storing the reproduced picture; and a motion vector coding means for performing the coding of the motion vectors. As above-mentioned, the fineness of the pattern of a block of the input television signal is considered as can be represented by a variance σ Thus, the motion compensated prediction interframe coding system according to the fourth aspect of the present invention can limit the amount of the generated codes and improve the picture of a block, which should be coded by using the intraframe coding method and has a fine pattern and of which the averaged value of gray levels of pixels is greater or equal to a predetermined value, by measuring the fineness of a pattern of each block of the luminance signal contained in the input television signal in terms of the variance σ Other features, objects and advantages of the present invention will become apparent from the following description of preferred embodiments with reference to the drawings in which like reference characters designate like or corresponding parts throughout several views, and in which: FIG. 1 is a schematic block diagram for showing a prior art motion compensated prediction interframe coding system; FIG. 2 is a diagram for illustrating the relation between a picture represented by an input television signal and a Macro Block; FIG. 3 is a schematic block diagram for showing a first motion compensated prediction interframe coding system embodying the present invention; FIG. 4 is a schematic block diagram for showing a second motion compensated prediction interframe coding system embodying the present invention; FIG. 5 is a schematic block diagram for showing a third motion compensated prediction interframe coding system embodying the present invention; FIG. 6 is a schematic block diagram for showing a fourth motion compensated prediction interframe coding system embodying the present invention; and FIG. 7 is a diagram for illustrating the relation between the quantization step sizes and threshold values. Hereinafter, the preferred embodiments of the present invention will be described in detail by referring to the accompanying drawings. First, a first embodiment of the present invention will be described hereinbelow with reference to FIG. In this figure, reference numeral Hereinafter, an operation of this motion compensated prediction interframe coding system having the above described arrangement will be described. First, television signals are converted into digital television signals by an analog-to-digital (A/D) conversion circuit (not shown). Then, pixels represented by the digital television signals are divided into blocks each of which is a rectangular array composed of M×N pixels arranged in M columns and N rows. Further, the digital television signals are input to the system from the input terminal A motion compensation portion Then, the intra-loop filtering portion Subsequently, the prediction error evaluating portion Hereunder, will be described a method for calculating the second quantization step size The variance calculating portion where M represents the number of columns (or pixels on a row) of a block; N the number of rows of the block; p(i,j) a gray level of a pixel at an address (i,j) in the block; and P the averaged value of the gray levels of the pixels of the block (see the following equation (4)). Further, the second quantization step size calculating portion As will be described below, the second quantization step size calculating portion {circle around (1)} In case where 0≦σ {circle around (2)} In case where th {circle around (3)} In case where th {circle around (4)} In case where th Thereby, for a block having a fine pattern or texture, the second quantization step size is made to become smaller than the first quantization step size. Then, the quantization portion Subsequently, the reproduced picture calculating portion As is apparent from the foregoing description, this embodiments calculates the second quantization step size from the first quantization step size as proportional to the fineness of the pattern of each block among a group of continuous blocks, each of which is quantized by using a quantization step size, and quantizes the prediction error of a block having a fine pattern by using the the second quantization step size. Thereby, the picture quality of the reproduced picture can be improved without degrading the fineness of the pattern or texture of the entire picture. Incidentally, the variance where p(i,j) denotes a gray level of a pixel at an address (i,j) in the block; and P the averaged value of the gray levels of the pixels of the block. Further, in this embodiment, the number of the classes is four as above described. However, a different number may be employed as the number of the classes. Additionally, in this embodiment, the second quantization step size Qstep is determined for each class as described in {circle around (1)}˜{circle around (4)}. Another method for determining the second quantization step size may employed as long as the second quantization step size can be determined by the method to be small for a block having a small variance. As described above, this embodiment first measures characteristics regarding the fineness of a pattern or texture of each block of picture, which is represented by an input television signal and is divided into a group of continuous blocks each quantized by using a quantization step size, and changes the quantization step size into a smaller one if quantizes a block having a picture pattern or texture finer than patterns of the other blocks. Thereby, an amount of generated codes can be limited but a motion compensated prediction frame coding operation can be performed without degrading the fineness of the texture of the original input picture. Further, the picture quality of the reproduced picture can be improved. Thus, practically, the present invention is very efficacious. Next, a second embodiment of the present invention will be described hereinbelow with reference to FIG. In this figure, reference numeral Hereinafter, an operation of this motion compensated prediction interframe coding system having the above described arrangement will be described. First, television signals are converted into digital television signals by an A/D conversion circuit (not shown). Then, pixels represented by the digital television signals are divided into blocks each of which is a rectangular array composed of M×N pixels arranged in M columns and N rows. Further, the digital television signals are input to the system from the input terminal A motion compensation portion The coding method selection portion Then, the intra-loop filtering portion Subsequently, the prediction error evaluating portion Thereafter, in case where the coding method selection signal Further, the orthogonal transform portion Then, the quantization portion (1) In case where a coding block is a block to be coded by the intraframe coding method: The second quantization step size calculating portion (2) In case where a coding block is a block to be coded by the interframe coding method: By effecting a process of steps (i) and (ii) as will be described later, a variance of gray levels of pixels of the coding block is first calculated. Then, the block is divided into classes according to the calculated variance. Further, the second quantization step size actually used in the quantization for each class is calculated from the first quantization step size. (i) The variance calculating portion where M represents the number of columns (or pixels on a row) of a block; N the number of rows of the block; p(i, j) a gray level of a pixel at an address (i, j) in the block; and P the averaged value of the gray levels of the pixels of the block (given by the following equation (7)). (ii) Further, the second quantization step size calculating portion The second quantization step size calculating portion {circle around (1)} In case where 0≦σ {circle around (2)} In case where th {circle around (3)} In case where th {circle around (4)} In case where th Thereby, the second quantization step size used for quantizing a block to be coded by using the interframe coding method is made to become smaller than the first quantization step size. Then, the quantization portion Further, another switching portion Then, the reproduced picture calculating portion Moreover, the prediction error coding portion Furthermore, the motion vector coding portion Thereafter, the multiplexer portion Subsequently, the code memory portion As is apparent from the foregoing description, this embodiments calculates the second quantization step size from the first quantization step size as proportional to the fineness of the pattern of each block among a group of continuous blocks, each of which is quantized by using a quantization step size, and quantizes the prediction error of a block having a fine pattern by using the the second quantization step size. Thereby, the picture quality of the entire reproduced picture can be improved without degrading the fineness of the pattern or texture of the picture. Incidentally, the variance where p(i, j) denotes a gray level of a pixel at an address (i, j) in the block; and P the averaged value of the gray levels of the pixels of the block. Further, in this embodiment, the number of the classes is four, as above described. However, a different number may be employed as the number of the classes so long as the second quantization step size can be determined to be small for a block having a small variance. As described above, this embodiment first measures characteristics regarding the fineness of a picture pattern or texture of each block of picture, which is represented by an input television signal and is divided into a group of continuous blocks each quantized by using a quantization step size, and changes the quantization step size into a smaller one when quantizing a block having a fine pattern or texture and to be coded by effecting the intra-frame coding method. Thereby, an amount of generated codes can be limited but a motion compensated prediction frame coding operation can be performed without degrading the fineness of the texture of the original input picture. Further, the picture quality of the reproduced picture can be improved. Therefore, practically, the present invention is very advantageous. Next, a third embodiment of the present invention will be described hereinbelow with reference to FIG. In this figure, reference numeral Hereinafter, an operation of this motion compensated prediction interframe coding system having the above described arrangement will be described. First, television signals are converted into digital television signals by an A/D conversion circuit (not shown). Then, pixels represented by the digital television signals are divided into blocks each of which is a rectangular array composed of M×N pixels arranged in M columns and N rows. Further, the digital television signals are input to the system from the input terminal A motion compensation predicting portion The coding method selection portion Then, the intra-loop filtering portion Subsequently, the prediction error evaluating portion Thereafter, in case where the coding method selection signal Further, the orthogonal transform portion Then, the quantization portion (1) In case where a coding block is a block to be coded by the intraframe coding method: The second quantization step size calculating portion (2) In case where a coding block is a block to be coded by the interframe coding method: By effecting a process of steps (I) and (II) as will be described later, a variance of gray levels of pixels of the coding block is first calculated. Then, the block is divided into classes according to the calculated variance. Moreover, the second quantization step size actually used in the quantization for each class is calculated from the first quantization step size. (I) The variance calculating portion Further, the variance σ where M represents the number of columns (or pixels on a row) of a block; N the number of rows of the block; and p(i,j) a gray level of a pixel at an address (i,j) in the block. (II) Further, the second quantization step size calculating portion The second quantization step size calculating portion {circle around (1)} In case where 0≦σ {circle around (2)} In case where th {circle around (3)} In case where th {circle around (4)} In case where th Thereby, the second quantization step size Then, the quantization portion Further, the switching portion Then, the reproduced picture calculating portion Moreover, the prediction error coding portion Furthermore, the motion vector coding portion Thereafter, the multiplexer portion Subsequently, the code memory portion As is apparent from the foregoing description, this embodiments calculates the second quantization step size from the first quantization step size as proportional to the fineness of the pattern of each block, which is coded by effecting the intra-frame coding method, among a group of continuous blocks, each of which is quantized by using a quantization step size, in case where the averaged value of the gray levels of pixels of an input picture is equal to or more than a predetermined constant value, and further quantizes the prediction error of a block having a fine pattern by using the second quantization step size. Thereby, the picture quality of the entire reproduced picture can be improved without degrading the fineness of the pattern or texture of the picture. Incidentally, the variance where p(i,j) denotes a gray level of a pixel at an address (i,j) in the block; and P the averaged value of the gray levels of the pixels of the block. Further, in this embodiment, it is determined depending on the averaged value of the gray levels of the pixels of the input picture whether or not the first quantization step size should be changed. However, such determination may be made depending on another measure which can appropriately represent characteristics of the gray levels of the pixels of the entire input picture. Furthermore, in this embodiment, the number of the classes is four, as above described, However, a different number may be employed as the number of the classes so long as the second quantization step size can be determined to be small for a block having a small variance. As described above, this embodiment first measures characteristics regarding the fineness of a picture pattern or texture of each block of picture, which is represented by an input television signal and is divided into a group of continuous blocks each quantized by using a quantization step size, and changes the quantization step size into a smaller one when quantizing a block having a fine pattern or texture and to be coded by effecting the intra-frame coding method, in case where the averaged value of the gray levels of the pixels of the entire input picture. Thereby, an amount of generated codes can be limited but the coding of a moving picture can be performed without degrading the fineness of the texture of the input or original picture. Further, the picture quality of the reproduced picture can be improved. Therefore, practically, the present invention is very advantageous. Especially, the picture quality of the refreshment frame can be substantially improved. Next, a fourth embodiment of the present invention will be described hereinbelow with reference to FIG. In this figure, reference numeral Hereinafter, an operation of this motion compensated prediction interframe coding system, having the above described arrangement will be described. First, television signals are converted into digital television signals by an A/D conversion circuit (not shown). Then, pixels represented by the digital television signals are divided into blocks each of which is a rectangular array composed of M×N pixels arranged in M columns and N rows. Further, the digital television signals are input to the system from the input terminal Then, the motion vector calculating portion A motion compensation predicting portion The coding method selection portion Then, the intra-loop filtering portion Subsequently, the prediction error evaluating portion Thereafter, in case where the coding method selection signal Further, the orthogonal transform portion Then, the quantization portion (1) In case where a coding block is a block to be coded by the intraframe coding method or is not a block represented by a luminance signal (hereunder referred to as a luminance signal block): The second quantization step size calculating portion (2) In case where a coding block is a block to be coded by the interframe coding method and is a luminance signal block: By effecting a process of steps (1) and (2) as will be described later, a variance of gray levels of pixels of the coding block is first calculated. Then, the block is divided into classes according to the calculated variance. Moreover, the second quantization step size actually used in the quantization for each class is calculated from the first quantization step size. (1) The variance calculating portion Further, the variance σ where M represents the number of columns (or pixels on a row) of a block; N the number of rows of the block; and p(i,j) a gray level of a pixel at an address (i,j) in the block. (2) Further, the second quantization step size calculating portion That is, the portion (1) In case where Qb≦thq, The threshold values th (2) In case where Qb>thq, The threshold values th
where a subscript i indicates an integer from 1 to 4; and “max” represents a maximum value of the quantization step size Qb. FIG. 7 shows the relation between the quantization step size Qb and the threshold value thi (i=1, 2, 3 and 4). Next, the portion {circle around (1)} In case where 0≦σ {circle around (2)} In case where th {circle around (3)} In case where th {circle around (4)} In case where th Thereby, the second quantization step size Further, the quantization portion Further, the switching portion Then, the reproduced picture calculating portion Moreover, the prediction error coding portion Furthermore, the motion vector coding portion Thereafter, the multiplexer portion Subsequently, the code memory portion As is apparent from the foregoing description, this embodiments calculates the second quantization step size from the first quantization step size as proportional to the fineness of the pattern of each block, which is coded by effecting the intra-frame coding method, among a group of continuous blocks, each of which is quantized by using a quantization step size, in case where the averaged value of the gray levels of pixels of an input picture is a predetermined constant value, and further quantizes the orthogonal transform coefficients by using the the second quantization step size. Thereby, the picture quality of the entire reproduced picture can be improved without degrading the fineness of the pattern or texture of the picture. Further, as above described, when the first quantization step size is more than a predetermined constant value, the system modifies the threshold values thi (i=1, 2, 3 and 4) for the classification, so that the number of blocks to be quantized by using a quantization step size smaller than the first quantization step size can be decreased and the amount of the code corresponding to the luminance signal can be reduced and the discoloration of the coding block, which is due to the fact that the amount of the code corresponding to the color difference signal is extremely small, can be prevented. Incidentally, the variance where p(i,j) denotes a gray level of a pixel at an address (i,j) in the block; and μ the averaged value of the gray levels of the pixels of the block. Further, in this embodiment, it is determined depending on the averaged value of the gray levels of the pixels of the input picture whether or not the first quantization step size should be changed. However, such determination may be made depending on another measure which can appropriately represent characteristics of the gray levels of the pixels of the entire input picture. Furthermore, in this embodiment, the number of the classes is also four, as above described. However, a classes is also four, as above described. However, a different number may be employed as the number of the classes so long as the second quantization step size can be determined to be small for a block having a small variance. Additionally, in this embodiment, the threshold values thi (i=1, 2, 3 and 4) for the classification are modified by the equation (14). Apparently, another method of modifying the threshold values for the classification may be employed so long as the threshold values for the classification modified by the method becomes smaller when the quantization step sizes become larger. While preferred embodiments of the present invention have been described above, it is to be understood that the present invention is not limited thereto and that other modifications will be apparent to those skilled in the art without departing from the spirit of the invention. The scope of the present invention, therefore, is to be determined solely by the appended claims. Patent Citations
Non-Patent Citations
Referenced by
Classifications
Rotate |