System and Method for Context-Based Adaptive Binary Arithematic Encoding and Decoding

• US 20090096643A1
• Filed: 08/11/2008
• Published: 04/16/2009
• Est. Priority Date: 10/16/2007
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Chang, Yuan-Teng
• Granted Patent: US 7,777,654 B2
• Time in Patent Office: Days
• US Class Current: 341/51
• CPC Class Codes: H03M 7/4006 Conversion to or from arithmetic code
• CPC Class Codes: H04N 19/426 using memory downsizing methods
• CPC Class Codes: H04N 19/436 using parallelised computational arrangements
• CPC Class Codes: H04N 19/91 Entropy coding, e.g. variable length coding [VLC] or arithmetic coding

Abstract

A system is disclosed for context-based adaptive binary arithmetic encoding and decoding. The system comprises a calculating device configured to calculate an index value for one of a first bin to be encoded and a second bin to be decoded, a memory device to store context models in cluster, wherein the index value for one of the first bin and the second bin is related to one of the context models, and a binary arithmetic unit configured to serve as one of an encoder for encoding the first bin based on the one context model and a decoder for decoding the second bin based on the one context model.

Claims

1. A system for context-based adaptive binary arithmetic encoding and decoding, the system comprising:

a calculating device configured to calculate an index value for one of a first bin to be encoded and a second bin to be decoded;

a memory device to store context models in cluster, wherein the index value for one of the first bin and the second bin is related to one of the context models; and

a binary arithmetic unit configured to serve as one of an encoder for encoding the first bin based on the one context model and a decoder for decoding the second bin based on the one context model.

2. The system of claim 1 further comprising a binarizer configured to binarize a syntax element into a bin string based on a value of the syntax element.

3. The system of claim 1, wherein the binary arithmetic unit includes at least one multiplexer controlled by a selector signal having a value related to one of an encoding status and a decoding status of one of the first bin and the second bin.

4. The system of claim 1, wherein the binary arithmetic unit includes a leading zero detector configured to detect the number (N) of leading zero in the binary form of a first output generated by the binary arithmetic unit, and a left shifter configured to shift the binary form of the first output to the left by N bit(s) if the first output is smaller than a predetermined value.

5. The system of claim 2 further comprising a bit generator configured to generate a bit string for each bin in the bin string.

6. The system of claim 1 further comprising:

a buffer to store a bin value for each bin of a syntax element; and

a de-binarizer configured to detect whether a decoding process for the syntax element is completed and output a value of the syntax element based on the bin value for the each bin when the decoding process for the syntax element is completed.

7. The system of claim 1 further comprising a first buffer to store information from a top block and a second buffer to store information from a left block, wherein the calculating device includes a first calculating module configured to calculate a first variable based on the information from the first buffer and the second buffer.

8. The system of claim 7, wherein the calculating device includes a second calculating module configured to calculate the index value based on the value of the first variable.

9. The system of claim 1, wherein the binary arithmetic unit is configured to update the one context model, and store an updated context model in the memory device.

10. The system of claim 9 further comprising:

a first lookup table configured to generate a first value of a second variable in response to the one context model;

a second lookup table configured to generate a second value of the second variable in response to the updated context model; and

a multiplexer to select one of the first value and the second value.

11. The system of claim 1, wherein the memory device includes a first register bank to store a context model cluster for a syntax element “

last_significant_coeff_flag”

(LSCF), and a second register bank to store context model clusters for other syntax elements.

12. The system of claim 1, wherein the memory device is configured to store context model clusters in a number of addresses, wherein context model clusters stored at each of the addresses satisfies:

#CMs (SE_i)+#CMs (SE_i+1)+ . . . +#CMs (SE_i+n)≦

P

where SE_i represents an i-th syntax element, #CMs (SE_i) represents the number of context models used in processing the SE_i, SE_i, SE_i+1, . . . , SE_i+n being arranged in an order of processing, and P represents the maximum number of sets of context models of a syntax element among all syntax elements.

13. A method of context-based adaptive binary arithmetic encoding and decoding in a pipeline flow, the method comprising:

reading context models related to a syntax element having a number of M bins, obtaining information from at least one neighbor block and calculating a variable based on the information from the at least one neighbor block for the syntax element at a first clock cycle;

calculating a first index value for a first bin of the syntax element based on the variable and loading one of the context models for the first bin based on the first index value at a second clock cycle;

processing the first bin in one of an encoding process and a decoding process based on the one context model related to the first bin and updating said one context model after processing the first bin at a third clock cycle;

calculating a second index value for a second bin of the syntax element based on the variable and loading one of the context models for the second bin based on the second index value at the third clock cycle; and

processing the second bin in the one of the encoding process and the decoding process based on the one context model related to the second bin and updating said one context model after processing the second bin at a fourth clock cycle.

14. The method of claim 13 further comprising:

calculating an i-th index value for an i-th bin of the syntax element based on the variable and loading one of the context models for the i-th bin based on the i-th index value at an (i+1)-th clock cycle; and

processing the i-th bin in the one of the encoding process and the decoding process based on the one context model related to the i-th bin and updating said context model after processing the i-th bin at an (i+2)-th clock cycle, i≦

M.

15. The method of claim 13 further comprising:

calculating an i-th index value for an i-th bin of the syntax element based on the variable and loading one of the context models for the i-th bin based on the i-th index value at an (i+1)-th clock cycle;

encoding the i-th bin and updating the one context model related to the i-th bin at an (i+2)-th clock cycle; and

generating a coded bit string for the i-th bin at an (i+3)-th clock cycle.

16. The method of claim 15 further comprising:

calculating an (i+1)-th index value for an (i+1)-th bin of the syntax element based on the variable and loading one of the context models for the (i+1)-th bin based on the (i+1)-th index value at the (i+1)-th clock cycle;

encoding the (i+1)-th bin and updating the one context model related to the (i+1)-th bin at the (i+3)-th clock cycle; and

generating a coded bit string for the (i+1)-th bin at an (i+4)-th clock cycle.

17. The method of claim 16 further comprising:

calculating an (i+2)-th index value for an (i+2)-th bin of the syntax element based on the variable and loading one of the context models for the (i+2)-th bin based on the (i+2)-th index value at the (i+2)-th clock cycle;

encoding the (i+2)-th bin and updating the one context model related to the (i+2)-th bin at the (i+4)-th clock cycle; and

generating a coded bit string for the (i+2)-th bin at an (i+5)-th clock cycle.

18. The method of claim 13 further comprising:

calculating an i-th index value for an i-th bin of the syntax element based on the variable and loading one of the context models for the i-th bin based on the i-th index value at an (i+1)-th clock cycle;

decoding the i-th bin and updating the one context model related to the i-th bin at an (i+2)-th clock cycle; and

storing a decoded bin value for the i-th bin and detecting whether the decoding process for the syntax element is completed at the (i+2)-th clock cycle.

19. The method of claim 18 further comprising:

calculating an (i+1)-th index value for an (i+1)-th bin of the syntax element based on the variable and loading one of the context models for the (i+1)-th bin based on the (i+1)-th index value at an (i+2)-th clock cycle;

decoding the (i+1)-th bin and updating the one context model related to the (i+1)-th bin at an (i+3)-th clock cycle; and

storing a decoded bin value for the (i+1)-th bin and detecting whether the decoding process for the syntax element is completed at the (i+3)-th clock cycle.

20. The method of claim 19 further comprising:

calculating an (i+2)-th index value for an (i+2)-th bin of the syntax element based on the variable and loading one of the context models for the (i+2)-th bin based on the (i+2)-th index value at an (i+3)-th clock cycle;

decoding the (i+2)-th bin and updating the one context model related to the (i+2)-th bin at an (i+4)-th clock cycle; and

storing a decoded bin value for the (i+2)-th bin and detecting whether the decoding process for the syntax element is completed at the (i+4)-th clock cycle.

21. The method of claim 20 further comprising:

calculating an M-th index value for an M-th bin of the syntax element based on the variable and loading one of the context models for the M-th bin based on the M-th index value at an (M+1)-th clock cycle;

decoding the M-th bin and updating the one context model related to the M-th bin at an (M+2)-th clock cycle; and

storing a decoded bin value for the M-th bin and detecting that the decoding process for the syntax element is completed at the (M+2)-th clock cycle.

22. The method of claim 21 further comprising:

debinarizing the syntax element and obtaining a value of the syntax element at the (M+2)-th clock cycle; and

writing back context models related to each of the i-th bin at an (M+3)-th clock cycle, i from 1 to M.

23. The method of claim 22 further comprising:

writing the value of the syntax element to a first buffer configured to store information from a top block and a second buffer configured to store information from a left block at the (M+3)-th clock cycle.

24. The method of claim 13 further comprising:

binarizing the syntax element into a bin string based on a value of the syntax element at the first clock cycle.

25. The method of claim 13 further comprising:

storing context model in cluster in a number of addresses of a memory device, wherein context model clusters stored at each of the addresses satisfies;

#CMs (SE_i)+#CMs (SE_i+1)+ . . . +#CMs (SE_i+n)≦

P

where SE_i represents an i-th syntax element, #CMs (SE_i) represents the number of context models used in processing the SE_i, SE_i, SE_i+1, . . . , SE_i+n being arranged in an order of processing, and P represents the maximum number of sets of context models of a syntax element among all syntax elements.

26. A method of context-based adaptive binary arithmetic encoding and decoding in a pipeline flow, the method comprising:

reading context models related to a first syntax element, obtaining information from at least one neighbor block and calculating a variable based on the information from the at least one neighbor block at a first clock cycle;

calculating a first index value for a first bin of the first syntax element based on the variable and loading one of the context models for the first bin of the first syntax element based on said first index value at a second clock cycle;

reading context models related to a second syntax element at the second clock cycle, wherein the first syntax element and the second syntax element occur in pair in data processing;

decoding the first bin of the first syntax element based on the one context model related to the first bin of the first syntax element and obtaining a value of the first bin of the first syntax element at a third clock cycle;

calculating a first index value for a first bin of the second syntax element based on the variable and loading one of the context models for the first bin of the second syntax element based on said first index value at the third clock cycle in response to one level of the value of the first bin of the first syntax element; and

calculating a second index value for a second bin of the first syntax element based on the variable and loading one of the context models for the second bin of the first syntax element based on said second index value at the third clock cycle in response to the other one level of the value of the first bin of the first syntax element.

27. The method of claim 26, wherein the first syntax element includes “

significant_coeff_flag” and

the second syntax element includes “

last_significant_coeff_flag”

.

28. The method of claim 26 further comprising:

calculating an i-th index value for an i-th bin of the first syntax element based on the variable and loading one of the context models for the i-th bin of the first syntax element based on said i-th index value at an (i+1) clock cycle;

decoding the i-th bin of the first syntax element based on the one context model related to the i-th bin of the first syntax element and obtaining a value of the i-th bin of the first syntax element at an (i+2) clock cycle;

calculating an i-th index value for an i-th bin of the second syntax element based on the variable and loading one of the context models for the i-th bin of the second syntax element based on said i-th index value at the (i+2) clock cycle in response to one level of the value of the i-th bin of the first syntax element; and

calculating an (i+1) index value for an (i+1) bin of the first syntax element based on the variable and loading one of the context models for the (i+1) bin of the first syntax element based on said (i+1) index value at the (i+2) clock cycle in response to the other one level of the value of the i-th bin of the first syntax element.

29. The method of claim 26 further comprising:

storing context model in cluster in a number of addresses of a memory device, wherein context model clusters stored at each of the addresses satisfies;

#CMs (SE_i)+#CMs (SE_i+1)+ . . . +#CMs (SE_i+n)<

P

where SE_i represents an i-th syntax element, #CMs (SE_i) represents the number of context models used in processing the SE_i, SE_i, SE_i+1, . . . , SE_i+n being arranged in an order of processing, and P represents the maximum number of sets of context models of a syntax element among all syntax elements.

30. The method of claim 26 further comprising:

assigning a first register bank to store a context model cluster for the second syntax element “

last_significant_coeff_flag”

(LSCF); and

assigning a second register bank to store context model clusters for other syntax elements including the first syntax element.