Video encoding control using non-exclusive content categories
View Patent ↗In some embodiments, macroblock-level encoding parameters are assigned to weighted Linear combinations of corresponding content-category-level encoding parameters. A macroblock quantization parameter (QP) modulation is set to a weighted linear combination of content category QP modulations. Content categories may identify potentially overlapping content types. The combination weights may be similarity measures describing macroblock similarities to content categories. A macroblock may be associated with multiple content categories, with different similarity levels for different content categories. A similarity measure for a given macroblock with respect to a content category may be defined as a number (between 0 and 8) of neighboring macroblocks that meet a similarity condition, provided the macroblock meets a qualification condition. The similarity condition may be computationally simpler than the qualification condition. Content-category-level statistics are generated by combining block-level statistics. Content-category-level statistics may be used in encoding subsequent frames.
1. A video encoding method comprising:
determining a plurality of weights for a video image block, each individual weight being an used as indicator of a similarity of the video image block to a distinct content category in a plurality of content categories, each content category in the plurality of content categories being associated with a content-category-level quantization precision indicator;
setting a quantization precision indicator for the video image block to a linear combination of corresponding content-category-level quantization precision indicators weighted according to the plurality of weights; and
encoding the video image block using the quantization precision indicator for the video image block.
2. The method of claim 1 , wherein the quantization precision indicator comprises a quantization parameter modulation, and wherein encoding the video image block comprises quantizing a set of frequency-domain transform coefficients according to a quantization parameter derived from the quantization parameter modulation.
3. The method of claim 1 , wherein said each weight characterizes a similarity of a set of neighbors of the video image block to the content category.
4. The method of claim 1 , wherein said each weight is proportional to a number of neighbors of the video image block that meet a similarity condition with respect to the content category.
5. The method of claim 1 , further comprising determining the plurality of quantization precision indicators by performing a content-category-level rate-distortion analysis in a ρ-domain.
6. The method of claim 5 , wherein each of the plurality of quantization precision indicators is a content category quantization parameter determined by interpolation from a set of quantization parameter-ρ correspondence values.
7. The method of claim 6 , wherein performing the rate-distortion analysis comprises employing a slope θ extracted from a set of content-category-level statistics computed for a set of frames preceding a frame of the video image block.
8. The method of claim 7 , further comprising extracting the set of content-category-level statistics by weighted linear combination from a set of block-level statistics for the set of the frame preceding the frame of the video image block.
9. A video encoding method comprising:
performing a weighted linear transformation between a set of block-level encoding parameters and a set of content-category level extensions of the encoding parameters, such that (1) each individual block-level encoding parameter is associated with a first linear combination of the content-category-level extensions in the set of content-category-level extensions of the encoding parameters according to a first plurality of weights, and (2) each individual content-category-level extension of the encoding parameter is associated with a second linear combination of the block-level encoding parameters in the set of block-level encoding parameters according to a second plurality of weights; and
encoding video data using the block-level encoding parameters.
10. A video encoding method comprising:
assigning a content category target rate for each of a plurality of content categories for a video frame;
assigning a content category rate slope θ r for said each of the plurality of content categories;
determining a fraction of frequency-domain transform zero coefficients for said each of the plurality of content categories from the rate slope θ r and the content category target rate;
assigning a quantization precision indicator for said each of the plurality of content categories by interpolating between known correspondences between quantization precision indicators and fraction of frequency-domain zero coefficient values; and
assigning a quantization precision indicator for a video data block in the video frame by linearly combining quantization precision indicators assigned for at least a subset of the plurality of content categories with corresponding weights characterizing a similarity of the video data block to the subset of the plurality of content categories.