Cavlc enhancements for svc cgs enhancement layer coding
AssigneeQUALCOMM Incorporated
InventorMarta Karczewicz
Life of this patent
Estimating the term…
Abstract
Context-adaptive variable length coding (CAVLC) for scalable video coding (SVC) coarse granular scalability (CGS) enhancement layer coding that utilizes the number of coefficients larger than one (NLrg1) as a syntax element. The coding of coeff_token is based on TotalCoeff and NLrg1 values. Levels are adaptively encoded based on NLrg1 syntax element total_zeros is encoded by adaptively selecting VLC tables based on an estimated number of a neighboring blocks' total_zeros.
Claims25
A method for context-based adaptive variable length coding, the method comprising:
- receiving transformed block image data;
- determining syntax elements from the block input data, the syntax elements including a TotalCoeffs value, a number of quantized transform coefficients having an absolute value larger than 1 (NLrg1) value, a total_zeros value, a run_before value, and a level value;
- and
- coding the transformed block image data in accordance with the syntax values.
A context-adaptive variable length coding (VLC) apparatus, comprising:
- syntax element determination units that determine a TotalCoeffs value, a number of quantized transform coefficients having an absolute value larger than 1 (NLrg1) value, a total_zeros value, a run_before value, and a level value;
- a coeff_token generation unit that encodes coeff_token in accordance with the NLrg1 value and the TotalCoeffs value;
- a level code generation unit that determines level coding in accordance with the number of coefficients that are larger than 1 in the remaining coefficients to be coded;
- a total_zeros generation unit that determined total_zeros encoding based an estimated total_zeros value determined from a neighboring blocks' total_zeros;
- a run_before generation unit that generates a variable-length code from the values of run_before and zeros_left;
- and
- a variable length combination unit that combines the output of the coeff_token generation unit, the level code generation unit, the total_zeros generation unit and the run_before generation unit.
A video encoder comprising:
- means for receiving transformed block image data;
- means for determining syntax elements from the block input data, the syntax elements including a TotalCoeffs value, a number of quantized transform coefficients having an absolute value larger than 1 (NLrg1) value, a total_zeros value, a run_before value, and a level value;
- and
- means for coding the transformed block image data in accordance with the syntax values.
A computer-readable medium comprising executable instructions to perform a method of context-based adaptive variable length coding, the method comprising:
- receiving transformed block image data;
- determining syntax elements from the block input data, the syntax elements including a TotalCoeffs value, a number of quantized transform coefficients having an absolute value larger than 1 (NLrg1) value, a total_zeros value, a run_before value, and a level value;
- and
- coding the transformed block image data in accordance with the syntax values.
Dependencies are inferred from the claim text.