Method and system for using motion prediction to equalize video quality across intra-coded frames
View Patent ↗In a video processing system, a method and system for using motion estimation to equalize video quality across intra frames are provided. Inter-prediction processing may be enabled for macroblocks in an “I” frame to generate temporal statistics. The inter-prediction processing may comprise motion estimation and motion compensation. Spatial statistics may be generated from the macroblocks in the “I” frame. A rate control may be determined based on the generated temporal and spatial statistics. The rate control may allocate a number of bits for encoding the “I” frame and a may allocate a portion of those bits to be used for encoding the current macroblock. The rate control may also be utilized to determine the quantization parameters for macroblocks in the “I” frame. Quantization parameters and intra-predicted current macroblocks may be utilized to generate “I” frame prediction residuals for encoding.
1. A method for video encoding, the method comprising:
inter-prediction processing of at least one current macroblock of an intra-coded frame in a video sequence;
generating temporal statistics based on said inter-predicted at least one current macroblock of said intra-coded frame;
generating spatial statistics based on said at least one current macroblock of said intra-coded frame and reference pixels that are from the same frame as said current macroblock;
determining a rate control based on said generated temporal statistics and said generated spatial statistics;
determining quantization parameters for said at least one current macroblock of said intra-coded frame based on said determined rate control; and
intra-prediction encoding said at least one current macroblock that was inter-prediction processed of said intra-coded frame, wherein intra-prediction encoding said at least one current macroblock of said intra-coded frame that was inter-prediction processed further comprises:
generating prediction residuals based on said determined quantization parameters and said at least one current macroblock of said intra-coded frame; and
encoding said prediction residuals based on said quantization parameters.
2. The method according to claim 1 , further comprising generating a signal that enables said inter-prediction processing of said at least one current macroblock of an intra-coded frame.
3. The method according to claim 1 , wherein said inter-prediction processing of said at least one current macroblock of said intra-coded frame comprises motion estimation and motion compensation operations.
4. The method according to claim 1 , further comprising generating prediction residuals to be encoded based on said determined quantization parameters and said intra-predicted at least one current macroblock of said intra-coded frame.
5. The method according to claim 1 , further comprising determining a prediction mode for said at least one current macroblock of said intra-coded frame.
6. The method according to claim 5 , wherein said determined prediction mode is an intra-prediction mode.
7. The method according to claim 5 , further comprising determining a type of motion block size for said determined prediction mode.
8. The method according to claim 1 , wherein said generated temporal statistics comprise average sum of absolute differences parameters, and average sum of squared difference parameters.
9. The method according to claim 1 , further comprising allocating a number of bits to said intra-coded picture based on said determined rate control.
10. The method according to claim 9 , further comprising allocating a portion of said number of bits to said at least one current macroblock of said intra-coded frame based on said determined rate control.
11. A system for video encoding, the system comprising:
circuitry for inter-prediction processing of at least one current macroblock of an intra-coded frame in a video sequence;
circuitry for generating temporal statistics based on said inter-predicted at least one current macroblock of said intra-coded frame;
circuitry for generating spatial statistics based on said at least one current macroblock of said intra-coded frame; and reference pixels that are from the same frame as said current macroblock;
circuitry for determining a rate control based on said generated temporal statistics and said generated spatial statistics;
circuitry for determining quantization parameters for said at least one current macroblock of said intra-coded frame based on said determined rate control; and
circuitry for intra-prediction encoding said at least one current macroblock that was inter-prediction processed of said intra-coded frame, wherein intra-prediction encoding further comprises:
circuitry for generating prediction residuals based on said determined quantization parameters and said at least one current macroblock of said intra-coded frame; and
circuitry for encoding said prediction residuals based on said quantization parameters.
12. The system according to claim 11 , further comprising circuitry for generating a signal that enables said inter-prediction processing of said at least one current macroblock of an intra-coded frame.
13. The system according to claim 11 , wherein said circuitry for inter-prediction processing of said at least one current macroblock of said intra-coded frame comprises circuitry for motion estimation and circuitry for motion compensation operations.
14. The system according to claim 11 , further comprising circuitry for generating prediction residuals to be encoded based on said determined quantization parameters and said intra-predicted at least one current macroblock of said intra-coded frame.
15. The system according to claim 11 , further comprising circuitry for determining a prediction mode for said at least one current macroblock of said intra-coded frame.
16. The system according to claim 15 , wherein said determined prediction mode is an intra-prediction mode.
17. The system according to claim 15 , further comprising circuitry for determining a type of motion block size for said determined prediction mode.
18. The system according to claim 11 , wherein said circuitry for generating spatial statistics comprises circuitry for generating local variance parameters and local dynamic range parameters.
19. The system according to claim 11 , wherein said circuitry for generating temporal statistics comprises circuitry for generating average sum of absolute differences parameters, average sum of squared difference parameters, motion vector consistency parameters, and prediction consistency parameters.
20. The system according to claim 11 , further comprising circuitry for allocating a number of bits to said intra-coded frame based on said determined rate control.
21. The system according to claim 20 , further comprising circuitry for allocating a portion of said number of bits to said at least one current macroblock of said intra-coded frame based on said determined rate control.
22. The method according to claim 1 , wherein all macroblocks of said I frame are intra-prediction encoded.
23. The system according to claim 11 , wherein all macroblocks of said I frame are intra-prediction encoded.
24. The method of claim 1 , wherein generating spatial statistics comprises subtracting the reference pixels that are from the same frame as said current macroblock from said current macroblock, and determining a variance of the differences of said reference pixels and said current macroblock.
25. The method of claim 1 , wherein generating spatial statistics comprises subtracting the reference pixels that are from the same frame as said current macroblock from said current macroblock, and determining a dynamic range of the differences of said reference pixels and said current macroblock.
26. The system of claim 11 , wherein generating spatial statistics comprises subtracting the reference pixels that are from the same frame as said current macroblock from said current macroblock, and determining a variance of the differences of said reference pixels and said current macroblock.
27. The system of claim 11 , wherein generating spatial statistics comprises subtracting the reference pixels that are from the same frame as said current macroblock from said current macroblock, and determining a dynamic range of the differences of said reference pixels and said current macroblock.