Video Encoder Multi-Reference Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face low coding efficiency when dealing with large residual signals between a to-be-coded picture and its reference picture, as they typically use only one reference frame for motion compensation prediction.
Innovation Solution
The method involves obtaining multiple motion vector fields (MVs) from a set of reference pictures within a video frame sequence, dividing the to-be-processed picture into sub-picture blocks, and using these MVs to determine motion compensation signals from multiple reference frames, thereby improving prediction residuals and coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If only one reference picture is used for motion compensation prediction, then the device complexity is low, but the coding efficiency deteriorates when residual is large
Solution Approach 1:
The patent divides the reference picture set into multiple reference pictures and further divides the current picture into multiple sub-picture blocks. Each sub-picture block can independently select from multiple reference pictures, allowing localized optimization of prediction accuracy without requiring global complexity increase.
Solution Approach 2:
The patent extends the single-reference-picture approach by adding a new dimension of multiple reference pictures. This allows the system to select the most appropriate reference picture for each sub-picture block based on motion characteristics, thereby improving coding efficiency without proportionally increasing device complexity.
2Productivity
If multiple reference pictures are used for motion compensation prediction, then the coding efficiency is improved, but the device complexity increases
Solution Approach 1:
By dividing the current picture into multiple sub-picture blocks, the patent enables independent motion estimation for each block. This segmentation allows the system to use multiple reference pictures only where necessary, improving coding efficiency while controlling the increase in processing complexity through localized application.
Solution Approach 2:
The patent applies multiple reference picture processing partially - not all sub-picture blocks necessarily use all reference pictures. The system selects the appropriate number of reference pictures based on the specific motion characteristics of each sub-picture block, avoiding unnecessary processing complexity while maintaining coding efficiency improvements.
3Measurement precision
If multiple sub-picture blocks are processed with different reference pictures, then the prediction accuracy is improved, but the processing time increases
Solution Approach 1:
The patent segments the picture into sub-picture blocks that can be processed independently and in parallel. This segmentation enables prediction accuracy improvement through localized reference picture selection while reducing overall processing time through parallel processing of multiple blocks simultaneously.
Data Source
AI summary
A video coding method includes obtaining a first motion vector field (MVF) set, obtaining a plurality of second MVFs according to the first MVF set and based on a preset vector operation method, determining a motion compensation prediction (MCP) signal of the target sub picture block from M MCP reference pictures according to the plurality of second MVFs and based on an MCP algorithm, obtaining a prediction residual of the target sub picture block according to a picture signal of the target sub picture block and the MCP signal of the target sub picture block, performing coding processing on the prediction residual of the target sub picture block to obtain a coded signal of the prediction residual of the target sub picture block, and sending the coded signal of the prediction residual of the target sub picture block to a peer decoder.