Image Decoding With Selective Prediction Flag Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image decoding technologies face inefficiencies in processing small chroma blocks, leading to increased overhead and processing delays, reduced throughput, and compromised compression performance due to restrictive chroma block sizes and limitations in parallel processing capabilities.
Innovation Solution
A method for decoding images that determines block sizes and prediction modes based on a predetermined threshold, allowing for efficient parsing of prediction flags and optimizing processing without unnecessary flag parsing for equal-sized blocks, thereby improving prediction image generation performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If chroma block size is restricted to small blocks (e.g., 2x2/4x4/2x4), then prediction processing is simplified, but overhead per pixel increases and processing delay increases reducing overall throughput
Solution Approach 1:
The patent applies dynamics by making the chroma block size adaptive rather than fixed. The chroma block size is dynamically determined based on the luma block size and prediction mode: when luma block size is large (e.g., 64x64) and intra prediction is used, chroma block size is reduced to 8x8 or 16x16 for simplified processing; when luma block size is small or inter prediction is used, chroma block size matches luma block size for better compression. This dynamic adjustment resolves the contradiction between processing simplicity and throughput.
Solution Approach 2:
The patent changes the parameter of chroma block size based on luma block size and prediction mode. Specifically, the chroma block size is set to be equal to luma block size when luma block size is 64x64 or less, or when inter prediction is used. When luma block size is greater than 64x64 and intra prediction is used, chroma block size is reduced to 8x8 or 16x16. This parameter change allows the system to optimize between processing simplicity and throughput based on actual encoding conditions.
2Device complexity
If single tree is used for luma and chroma, then structure is simplified, but when chroma block size is restricted, luma block cannot be split resulting in increased size and reduced compression performance
Solution Approach 1:
The patent applies local quality by allowing different block size restrictions for luma and chroma components. While a single tree structure is used for both components, the patent enables independent control of block size for each component based on their specific requirements. Luma blocks can be split freely for compression, while chroma blocks are restricted to specific sizes (8x8 or 16x16) when luma blocks are large and intra prediction is used. This local differentiation maintains compression performance while simplifying the overall tree structure.
Solution Approach 2:
The patent segments the block size control into different rules for luma and chroma components. Although a single tree structure is used, the patent applies segmentation by allowing luma blocks to be split into smaller blocks for compression, while restricting chroma blocks to specific sizes. This segmentation of control rules resolves the contradiction between structural simplicity and compression performance.
3Measurement precision
If prediction mode flag is parsed for all blocks, then prediction accuracy is improved, but processing time increases especially for small chroma blocks
Solution Approach 1:
The patent applies partial action by parsing the prediction mode flag selectively rather than for all blocks. Specifically, the prediction mode flag is parsed only when the chroma block size is greater than 8x8 or when inter prediction is used. When chroma block size is 8x8 or 16x16 and intra prediction is used, the prediction mode flag is not parsed, as the prediction mode can be derived from the luma block information. This partial parsing reduces processing time while maintaining prediction accuracy.
Solution Approach 2:
The patent applies self-service by allowing the system to automatically determine prediction mode without explicit flag parsing for certain cases. When chroma block size is small (8x8 or 16x16) and intra prediction is used, the prediction mode is automatically set based on the luma block prediction mode, eliminating the need for explicit flag parsing. This self-determination mechanism reduces processing time while maintaining prediction accuracy.
Data Source
AI summary
A method of decoding a picture from a bitstream by an electronic device is provided. A block size of a block unit in the picture is determined. The block size of the block unit is compared to a predetermined size. When the block size of the block unit is greater than the predetermined size, a prediction mode of the block unit is determined by parsing a prediction mode flag of the block unit from the bitstream. When the block size of the block unit is equal to the predetermined size, the prediction mode of the block unit is determined without parsing the prediction mode flag of the block unit. The block unit is decoded based on the prediction mode.