Image encoding apparatus, image encoding method, and image encoding program

• US 8,396,311 B2
• Filed: 04/28/2008
• Issued: 03/12/2013
• Est. Priority Date: 05/21/2007
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Chono, Keiichi
• Primary Examiner(s): Chen, Wenpeng
• Assistant Examiner(s): Niu, Feng
• Application Number:
  US12/601,484
• Publication Number:
  US 20100172593A1
• Time in Patent Office: 1,779 Days
• Field of Search: None
• US Class Current: 382/239
• CPC Class Codes: H03M 7/4006 Conversion to or from arithmetic code
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/13 Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
• CPC Class Codes: H04N 19/14 Coding unit complexity, e.g. amount of activity or edge presence estimation H04N19/146 takes precedence
• CPC Class Codes: H04N 19/149 by estimating the code amount by means of a model, e.g. mathematical model or statistical model
• CPC Class Codes: H04N 19/15 by monitoring actual compressed data size at the memory before deciding storage at the transmission buffer
• CPC Class Codes: H04N 19/152 by measuring the fullness of the transmission buffer
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

An image encoding apparatus, an image encoding method, and an image encoding program capable of minimizing image degradation, controlling a code amount in units smaller than a picture, and ensuring that encoding of the picture is completed within a certain time period is provided. A symbol number estimating device estimates the total bin number of a plurality of macro blocks constituting a picture. An entropy encoding selector outputs the entropy encoding mode selecting signal to an entropy encoder using the inputted bin number, in order to select one of a CABAC device or a VLC device. A CPU performs image encoding by executing a control program stored in a program storing portion.

Claims

1. An image encoding apparatus comprising:

a code data generator to transform and quantize an image block to generate code data;

a variable length code encoder to encode the code data into bit-string data on the basis of applying variable length encoding;

an arithmetic encoder to encode the code data into bit-string data on the basis of converting the code data into binary symbols and applying binary arithmetic encoding;

a binary symbol number estimator to estimate the total number of the binary symbols that are converted from the code data generated from a plurality of image blocks of a target picture to be encoded; and

an encoding selector to select, for the encoding of the code data, either the variable length code encoder or the arithmetic encoder on the basis of the total number estimated by the binary symbol number estimator, wherein

the binary symbol number estimator;

obtains, for the plurality of image blocks of the target picture to be encoded, any one of a sum of activities, a sum of standard deviations, or a sum of variances, and

estimates the total number of the binary symbols by selecting a larger one of a first value and a second value,

the first value is a value obtained by dividing an estimated bin number complexity of the target picture to be encoded by a quantization step size obtained by dividing a product of a picture rate and an average quantizing step size of macro blocks of a lastly encoded picture associated with the same picture type as that of the target picture to be encoded by an assigned picture rate of the target picture to be encoded, wherein the estimated bin number complexity is estimated according to the equation epic_bin_x=α

×

pic_W+β

, wherein;

epic_bin_x is the estimated bin number complexity of the target picture to be encoded,

pic_W is the any one of a sum of activities, a sum of standard deviations, or a sum of variances,

α

is a gradient of a regression line of I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, the I picture bin number complexity being, a product of a bin number and an average quantizing step size of macro blocks of an I picture,

β

is an intercept of the regression line of the I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, and

the second value is a product of an assigned picture rate of the target picture to be encoded and a symbol-to-bit ratio, wherein the symbol-to-bit ratio is a value obtained statistically from a plurality of pictures having a same type as that of the target picture to be encoded or a value obtained from a lastly encoded picture having a same picture type as that of the target picture to be encoded.

2. The image encoding apparatus according to claim 1 wherein the binary symbol number estimator estimates, on the basis of the total number, a quantization step size corresponding to the code data amount that the capability of the variable length code encoder meets, and a quantization step size corresponding to the code data amount that the capability of the arithmetic encoder meets;

and

the encoding selector selects either the variable length code encoder or the arithmetic encoder that is associated with the smaller estimated quantization step size.

3. The image encoding apparatus according to claim 1 further comprising a threshold comparator to compare the total binary symbol number estimated by the binary symbol number estimator with a predetermined threshold value, wherein

the encoding selector selects the variable length code encoder if the threshold comparator determines that the total number is larger than the predetermined threshold value, and selects the arithmetic encoder otherwise.

4. The image encoding apparatus according to claim 3 wherein the threshold value is a value calculated from the maximum capable binary symbol number that the arithmetic encoder encodes within the encoding time of the plurality of image blocks and from the compression ratio between the binary symbol number and the number of the bit-string data of the arithmetic encoder.

5. An image encoding method comprising:

transforming and quantizing an image block to generate code data;

estimating the total number of the binary symbols that are converted from the code data generated from a plurality of image blocks of a target picture to be encoded; and

selecting the encoding method of the code data from variable length encoding and arithmetic encoding on the basis of the total number estimated, the variable length encoding being adapted to encode the code data into bit-string data on the basis of applying variable length encoding, and the arithmetic encoding being adapted to encode the code data into bit-string data on the basis of converting the code data into binary symbols and applying binary arithmetic encoding, wherein

the estimating the total number of the binary symbols comprises;

obtaining, for the plurality of image blocks of the target picture to be encoded, any one of a sum of activities, a sum of standard deviations, or a sum of variances, and

estimating the total number of the binary symbols by selecting a larger one of a first value and a second value,

the first value is a value obtained by an estimated bin number complexity of the target picture to be encoded by a quantization step size obtained by dividing a product of a picture rate and an average quantizing step size of macro blocks of a lastly encoded picture associated with the same picture type as that of the target picture to be encoded by an assigned picture rate of the target picture to be encoded, wherein the estimated bin number complexity is estimated according to the equation epic_bin_x=α

×

pic_W+β

, wherein;

epic_bin_x is the estimated bin number complexity of the target picture to be encoded,

pic_W is the any one of a sum of activities, a sum of standard deviations, or a sum of variances,

α

is a gradient of a regression line of I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, the I picture bin number complexity being a product of a bin number and an average quantizing step size of macro blocks of an I picture,

β

is an intercept of the regression line of the I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, and

the second value is a product of an assigned picture rate of the target picture to be encoded and a symbol-to-bit ratio, wherein the symbol-to-bit ratio is a value obtained statistically from a plurality of pictures having a same type as that of the target picture to be encoded or a value obtained from a lastly encoded picture having a same picture type as that of the target picture to be encoded.

6. The image encoding method according to claim 5 further estimating, on the basis of the total number, a quantization step size corresponding to the code data amount that the capability of the variable length encoding method meets, and a quantization step size corresponding to the code data amount that the capability of the arithmetic encoding method meets;

and

when selecting the encoding method of the code data, selecting either the variable length encoding method or the arithmetic encoding method that is associated with the smaller estimated quantization step size.

7. The image encoding method according to claim 5 further comparing the total binary symbol number estimated with a predetermined threshold value, and

when selecting the encoding method of the code data, selecting the variable length encoding method if the total number is determined larger than the predetermined threshold value, and selecting the arithmetic encoding method otherwise.

8. The image encoding method according to claim 7 wherein the threshold value is a value calculated from the maximum capable binary symbol number that the arithmetic method encodes within the encoding time of the plurality of image blocks and from the compression ratio between the binary symbol number and the number of the bit-string data of the arithmetic method.

9. A non-transitory computer-readable recording medium recording an image encoding program to instruct a computer to execute:

a code data generating process to transform and quantize an image block to generate code data;

a binary symbol number estimating process to estimate the total number of the binary symbols that are converted from the code data generated from a plurality of image blocks; and

an encoding selecting process to select the encoding method of the code data from variable length encoding and arithmetic encoding on the basis of the total number estimated by the binary symbol number estimating process, the variable length encoding being adapted to encode the code data into bit-string data on the basis of applying variable length encoding, and the arithmetic encoding being adapted to encode the code data into bit-string data on the basis of converting the code data into binary symbols and applying binary arithmetic encoding, wherein

the binary symbol number estimating process;

obtains, for the plurality of image blocks of the target picture to be encoded, any one of a sum of activities, a sum of standard deviations, or a sum of variances, and

estimates the total number of the binary symbols by selecting a larger one of a first value and a second value,

the first value is a value obtained by dividing an estimated bin number complexity of the target picture to be encoded by a quantization step size obtained by dividing a product of a picture rate and an average quantizing step size of macro blocks of a lastly encoded picture associated with the same picture type as that of the target picture to be encoded by an assigned picture rate of the target picture to be encoded, wherein the estimated bin number complexity is estimated according to the equation epic_bin_x=α

×

pic_W+β

, wherein;

epic_bin_x is the estimated bin number complexity of the target picture to be encoded,

pic_W is the any one of a sum of activities, a sum of standard deviations, or a sum of variances,

α

is a gradient of a regression line of I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, the I picture bin number complexity being a product of a bin number and an average quantizing step size of macro blocks of an I picture,

β

is an intercept of the regression line of the I picture bin number complexity for the one of a sum of activities, a sum of standard deviations, or a sum of variances, and

the second value is a product of an assigned picture rate of the target picture to be encoded and a symbol-to-bit ratio, wherein the symbol-to-bit ratio is a value obtained statistically from a plurality of pictures having a same type as that of the target picture to be encoded or a value obtained from a lastly encoded picture having a same picture type as that of the target picture to be encoded.

10. The computer-readable recording medium according to claim 9 wherein the binary symbol number estimating process estimates, on the basis of the total number, a quantization step size corresponding to the code data amount that the capability of by using the variable length encoding method meets, and a quantization step size corresponding to the code data amount that the capability of the arithmetic encoding method meets;

and

the encoding selecting process selects either the variable length encoding method or the arithmetic encoding method that is associated with the smaller estimated quantization step size.

11. The computer-readable recording medium according to claim 9, wherein the program further instructing a computer to execute a threshold comparing process to compare the total binary symbol number estimated by the binary symbol number estimating process with a predetermined threshold value, wherein

the encoding selecting process selects the variable length encoding method if the threshold comparing process determines that the total number is larger than the predetermined threshold value, and selects the arithmetic encoding method otherwise.

12. The computer-readable recording medium according to claim 11 wherein the threshold value is a value calculated from the maximum capable binary symbol number that the arithmetic method encodes within the encoding time of the plurality of image blocks and from the compression ratio between the binary symbol number and the number of the bit-string data of the arithmetic method.