Weighted Prediction Signaling for Lower Video Coding Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution, high-quality images and videos, particularly in virtual reality and augmented reality, necessitates a more efficient compression technique to reduce transmission and storage costs while effectively handling diverse image/video characteristics.
Innovation Solution
A method and apparatus for enhancing image/video coding efficiency by signaling a prediction weighted table syntax, reducing signaling overhead, and minimizing the number of bits used for weighted prediction through flags and syntax elements in the bitstream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high resolution and high quality image/video are transmitted or stored using existing methods, then image/video quality is maintained, but transmission and storage costs are readily increased
Solution Approach 1:
The patent extracts only the essential visual information needed for high-quality reconstruction by using advanced prediction techniques (intra-picture and inter-picture prediction) to separate important detail information from redundant data, transmitting only the extracted essential components
Solution Approach 2:
The video coding apparatus performs multiple functions including prediction, transformation, quantization, and entropy coding within a unified framework that handles both intra-picture and inter-picture coding modes, achieving efficient compression across different video characteristics
2Measurement precision
If detailed weighted prediction information is signaled for every block, then prediction accuracy is improved, but signaling overhead is increased
Solution Approach 1:
The patent applies weighted prediction selectively rather than universally - using prediction weight tables and offset tables only when beneficial, with flags (cbf_luma_weight_10_flag, cbf_luma_offset_10_flag, etc.) controlling whether detailed weighted prediction information is signaled for each block, thereby avoiding unnecessary signaling overhead
Solution Approach 2:
The patent applies different prediction strategies to different regions of the video data - using advanced weighted prediction with prediction weight tables for blocks where it provides significant improvement, while using simpler prediction methods for blocks where advanced techniques would not provide substantial benefit
3Productivity
If advanced prediction techniques are used for all blocks, then compression efficiency is improved, but processing complexity is increased
Solution Approach 1:
The patent dynamically selects between different prediction modes (intra-picture, inter-picture, skip mode) and adjusts the use of advanced weighted prediction techniques based on block characteristics, reference picture availability, and coding conditions, allowing the system to adapt processing complexity to actual needs
Solution Approach 2:
The patent divides the video data into different block types and processing stages, applying advanced prediction techniques only to specific block types where they provide benefit, while using simpler methods for other blocks, thereby segmenting the processing complexity across different data regions
Data Source
AI summary
A video decoding method performed by a video decoding device may include: parsing a first flag and a second flag from a bitstream, wherein the first flag is related to whether weighted prediction is applied to slices referring to a picture parameter set within the bitstream, and the second flag is related to whether information relating to the weighted prediction exists in a picture header within the bitstream; parsing a prediction weighted table syntax from the picture header based on the first flag and the second flag; generating prediction samples for a current block in a current picture based on syntax elements within the prediction weighted table syntax; generating residual samples based on residual information obtained from the bitstream; and generating reconstructed samples based on the prediction samples and the residual samples, wherein the second flag is parsed from the picture parameter set based on the first flag.