Encoding device and method for encoding digital image improved by requantization of a predetermined subbit plane

• US 6,774,824 B2
• Filed: 07/09/2003
• Issued: 08/10/2004
• Est. Priority Date: 07/10/2002
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Sakaguchi, Tomoyuki
• Primary Examiner(s): Tokar, Michael
• Assistant Examiner(s): Mai, Lam T.
• Application Number:
  US10/616,068
• Publication Number:
  US 20040032989A1
• Time in Patent Office: 398 Days
• Field of Search: 341/50, 341/64, 382/239, 382/251, 382/232, 382/252, 709/203, 709/217
• US Class Current: 341/67
• 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 19/63 using sub-band based transform, e.g. wavelets

Abstract

There is provided a digital image encoding device which can enhance image compression performance while maintaining an image quality. When coefficient bits are decomposed/aligned in four encoded paths for each context by a procedure called coefficient bit modeling of a JPEG 2000 encoding system, if an appearance frequency of “1” is low in subbit plane coefficients of a clean-up path of a low-order bit plane n (bit plane n is lower in order than bit plane np represented by threshold value np), coefficients of subbit planes of this clean-up path are all reset to “0.” Thus, a run length used when a symbol and a context are generated is made longer and, as a result, the amount of codes after arithmetic encoding at a subsequent stage is reduced while an image quality deterioration is suppressed.

Claims

1. A digital image encoding device comprising:

conversion means for converting digital image data into a coefficient based on a spatial frequency;

quantization means for quantizing the coefficient obtained by said conversion means; and

coefficient bit modeling means for carrying out coefficient bit modeling by using said coefficient quantized by said quantization means,

wherein said digital image data is divided into a plurality of subbands in said conversion means,

wherein said coefficient bit modeling means comprises;

code block division means for dividing each of said subbands into a plurality of code blocks;

bit plane decomposition means for decomposing each of coefficients of said code blocks obtained by said code block division means into a plurality of bit planes for each bit weight;

subbit plane decomposition means for decomposing each of said bit planes obtained by said bit plane decomposition means into a plurality of subbit planes based on usefulness; and

predetermined data generation means for generating predetermined data in accordance with said subbit planes obtained by said subbit plane decomposition means, and

requantization means for requantizing a predetermined subbit plane among said subbit planes obtained by said subbit plane decomposition means and passing data after said requantization to said predetermined data generation means.

2. The digital image encoding device according to claim 1, wherein said requantization means carries out said requantization, provided that the following three conditions are satisfied;

a first condition where a currently processed subband is not a band of a lowest frequency compared with other subbands,

a second condition where a currently processed bit plane has weight lighter than predetermined weight, and

a third condition where said number of “

1”

in a currently processed subbit plane is equal to or lower than a predetermined value.

3. The digital image encoding device according to claim 2, wherein said requantization means resets all currently processed subbit planes to “

0”

, when said three conditions are satisfied.

4. The digital image encoding device according to claim 2, wherein said requantization means resets all currently processed subbit planes and subbit planes lower in order than said subbit planes to “

0”

, when said three conditions are satisfied.

5. The digital image encoding device according to claim 2, wherein said requantization means cancels processing of a currently processed code block and carries out processing of a next code block, when said three conditions are satisfied.

6. The digital image encoding device according to claim 2, wherein said predetermined weight of said second condition and said predetermined value of said number of “

1”

of said third condition can be selected for every subband and in accordance with a decomposition level of said conversion means.

7. A digital image encoding means comprises:

conversion step of converting digital image of data into a coefficient based on a spatial frequency;

quantization step of quantizing said coefficient obtained by said conversion step; and

coefficient bit modeling step of carrying out coefficient bit modeling by using said coefficient quantized by said quantization step,

wherein said digital image data is divided into a plurality of subbands in said conversion step,

wherein said coefficient bit modeling step comprises;

code block division step of dividing each of said subbands into a plurality of code blocks;

bit plane decomposition step of decomposing each of coefficients of said code blocks obtained by said code block division means into a plurality of bit planes for each bit weight;

subbit plane decomposition step of decomposing each of said bit planes obtained by said bit plane decomposition step into a plurality of subbit planes based on usefulness; and

predetermined data generation step of generating predetermined data in accordance with said subbit planes obtained by said subbit plane decomposition step, and

requantization step of requantizing a predetermined subbit plane among said subbit planes obtained by said subbit plane decomposition step and passing data after said requantization to said predetermined data generating step.

8. The digital image encoding method according to claim 7, wherein said requantization step carries out said requantization, provided that the following three conditions are satisfied;

a first condition where a currently processed subband is not a band of a lowest frequency compared with other subbands,

a second condition where a currently processed bit plane has weight lighter than predetermined weight, and

a third condition where said number of “

1”

in a currently processed subbit plane is equal to or lower than a predetermined value.

9. The digital image encoding method according to claim 8, wherein said requantization step resets all currently processed subbit planes to “

0”

, when said three conditions are satisfied.

10. The digital image encoding method according to claim 8, wherein said requantization step resets all currently processed subbit planes and subbit planes lower in order than said subbit planes to “

0”

, when said three conditions are satisfied.

11. The digital image encoding method according to claim 8, wherein said requantization step cancels processing of a currently processed code block and carries out processing of a next code block, when said three conditions are satisfied.

12. The digital image encoding method according to claim 8, wherein said predetermined weight of said second condition and said predetermined value of said number of “

1”

of said third condition can be selected for every subband and in accordance with a decomposition level of said conversion step.

13. A program for causing a computer to perform a digital image encoding method, said method comprising:

conversion step of converting digital image data into a coefficient based on a spatial frequency;

quantization step of quantizing said coefficient obtained by said conversion step; and

coefficient bit modeling step of carrying out coefficient bit modeling by using said coefficient quantized by said quantization step,

wherein said digital image data is divided into a plurality of subbands in said conversion step,

wherein said coefficient bit modeling step comprises;

code block division step of dividing each of said subbands into a plurality of code blocks;

bit plane decomposition step of decomposing each of coefficients of said code blocks obtained by said code block division means into a plurality of bit planes for each bit weight;

subbit plane decomposition step of decomposing each of said bit planes obtained by said bit plane decomposition step into a plurality of subbit planes based on usefulness; and

predetermined data generation step of generating predetermined data in accordance with said subbit planes obtained by said subbit plane decomposition step, and

requantization step of requantizing a predetermined subbit plane among said subbit planes obtained by said subbit plane decomposition step and passing data after said requantization to said predetermined data generation step.

14. The program according to claim 13, wherein said requantization step carries out said requantization, provided that the following three conditions are satisfied;

a first condition where a currently processed subband is not a band of a lowest frequency compared with other subbands,

a second condition where a currently processed bit plane has weight lighter than predetermined weight, and

a third condition where said number of “

1”

in a currently processed subbit plane is equal to or lower than a predetermined value.

15. The program according to claim 14, wherein said requantization step resets all currently processed subbit planes to “

0”

, when said three conditions are satisfied.

16. The program according to claim 14, wherein said requantization step resets all currently processed subbit planes and subbit planes lower in order than said subbit planes to “

0”

, when said three conditions are satisfied.

17. The program according to claim 14, wherein said requantization step cancels processing of a currently processed code block and carries out processing of a next code block, when said three conditions are satisfied.

18. The program according to claim 14, wherein said predetermined weight of said second condition and said predetermined value of said number of “

1”

of said third condition can be selected for every subband and in accordance with a decomposition level of said conversion step.