Inter Prediction Mode Signaling for Lower Compression Bit Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high resolution and quality images leads to increased data volume, resulting in higher transmission and storage costs, and existing video compression techniques face challenges with efficient encoding and decoding of diverse inter prediction modes, which contribute to compression bit overhead.
Innovation Solution
An image encoding/decoding method that includes decoding/encoding subblock merge mode indicators and regular merge mode indicators based on block size, type, and combined intra prediction use, allowing for efficient derivation of inter prediction information using MMVD or geometric partitioning modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If various inter prediction modes are diversified to improve prediction performance, then prediction accuracy is improved, but compression bit overhead increases
Solution Approach 1:
The current block is divided into subblocks, and different merge modes can be applied to different subblocks. This segmentation allows the encoder to use detailed prediction modes only where necessary while using simpler modes elsewhere, improving prediction accuracy in critical regions while controlling overall bit overhead.
Solution Approach 2:
The patent applies complex inter prediction modes selectively rather than universally. By using subblock-level merge modes only for blocks that benefit from them while using regular merge modes for others, the system achieves partial application of advanced techniques, balancing prediction performance improvement with bit rate control.
2Measurement precision
If subblock merge mode is used to improve prediction accuracy, then prediction performance is improved, but encoding complexity increases
Solution Approach 1:
The patent dynamically selects between regular merge mode and subblock merge mode based on block characteristics such as size, slice type, and prediction mode. This dynamic adaptation allows the encoder to increase complexity only when beneficial for prediction accuracy, rather than always using the most complex mode.
Solution Approach 2:
Different merge modes are applied to different regions (subblocks) of the current block based on local characteristics. This allows the encoder to apply higher complexity processing only to specific subblocks where it provides benefit, rather than uniformly increasing complexity across the entire block.
3Measurement precision
If regular merge mode indicator is always decoded to ensure accurate mode selection, then mode selection accuracy is improved, but bit overhead increases
Solution Approach 1:
The regular merge mode indicator is decoded selectively rather than always. The decoder determines whether to decode it based on the subblock merge mode indicator value and other conditions, applying full mode selection accuracy only when necessary while reducing overhead in other cases.
Data Source
AI summary
Disclosed herein is an image encoding/decoding method and apparatus. An image decoding method includes decoding a subblock merge mode indicator of a current block, decoding a regular merge mode indicator of the current block, if the subblock merge mode indicator does not indicate a subblock merge mode, and decoding subblock merge mode information of the current block, if the subblock merge mode indicator indicates a subblock merge mode.