Intra-prediction mode coding using quantized transform coefficients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression techniques face inefficiencies due to high bit rates required for encoding intra-prediction modes, especially when a large number of modes are available, leading to over-saturation issues where the number of bits needed to encode a mode exceeds the number of pixels in a block.
Innovation Solution
The proposed solution involves using information from residual blocks, specifically quantized transform coefficients, to improve the entropy coding of intra-prediction modes, reducing bit rates by determining a context for encoding based on the distribution of coefficients within the block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large number of intra-prediction modes are available to improve encoding accuracy, then the precision of video reconstruction is improved, but the bit rate required to encode the intra-prediction mode increases excessively
Solution Approach 1:
The patent changes the parameter of intra-prediction mode selection by using quantized transform coefficient information to determine a category for the intra-prediction mode. This parameter change allows the encoder to select from a reduced set of categories rather than all possible modes, thereby reducing the bit rate while maintaining adequate reconstruction precision through context-adaptive binary arithmetic coding (CABAC).
Solution Approach 2:
The patent segments the intra-prediction mode selection process into two stages: first determining a category based on quantized transform coefficient characteristics, then selecting the specific mode within that category. This segmentation reduces the overall search space and encoding complexity, allowing fewer bits to be allocated while preserving essential prediction accuracy.
2Productivity
If context-adaptive binary arithmetic coding (CABAC) is used to improve encoding efficiency, then the compression ratio is improved, but the computational complexity of encoding and decoding increases
Solution Approach 1:
The patent applies preliminary action by determining the category of intra-prediction mode based on quantized transform coefficient information before performing the actual CABAC encoding. This preliminary classification prepares the data in a way that simplifies subsequent CABAC processing, allowing the complex arithmetic coding to operate on a more structured and predictable input, thereby reducing the overall computational burden.
3Measurement precision
If the number of bits allocated to encode intra-prediction mode is increased to improve prediction accuracy, then the video quality is improved, but the total bit rate of the encoded video stream increases
Solution Approach 1:
The patent changes the encoding parameter from directly encoding the full intra-prediction mode to first encoding a category derived from quantized transform coefficient characteristics. This parameter transformation reduces the number of bits required while maintaining prediction accuracy through the relationship between coefficient distribution and optimal prediction modes.
Solution Approach 2:
The patent employs feedback by using quantized transform coefficient information (which reflects the actual content characteristics) to guide the selection and encoding of intra-prediction mode categories. This feedback mechanism ensures that the encoding process adapts to the local content characteristics, achieving high prediction accuracy with minimal bit rate allocation.
Data Source
AI summary
Coding a current block using an intra-prediction mode are disclosed. A method includes receiving a block of quantized transform coefficients, determining a context for coding the intra-prediction mode using the quantized transform coefficients, and coding the intra-prediction mode using the context. An apparatus, including a memory and a processor, for encoding and decoding a current block using an intra-prediction mode according is also disclosed. The processor is configured to execute instructions stored in the memory to decode, from an encoded bitstream, a block of quantized transform coefficients, determine a category for the intra-prediction mode using the quantized transform coefficients, and decode, from the encoded bitstream, the current block using the category.