Video Block Prediction Using a Frozen HMVP Candidate List
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding frameworks face challenges in parallel processing of image blocks due to dependencies in historical motion vector prediction candidate lists, which hinder efficient encoding and decoding.
Innovation Solution
A method and device that utilize a historical motion information candidate list for multiple image blocks, maintaining its consistency during prediction without updates, allowing parallel processing by using motion information from previous blocks to update the list after specific regions are encoded or decoded.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the historical motion information candidate list is updated for each image block during prediction, then the prediction accuracy is improved, but the processing time increases and parallel processing becomes difficult
Solution Approach 1:
The current frame is divided into multiple image blocks that are processed independently. By segmenting the frame into discrete blocks and using a shared historical motion information candidate list for all blocks, the patent enables parallel processing while maintaining prediction accuracy. Each block can be encoded or decoded independently using the same candidate list, eliminating sequential dependencies.
Solution Approach 2:
The historical motion information candidate list is prepared in advance and maintained consistently throughout the prediction process for all image blocks. This preliminary preparation allows multiple blocks to utilize the same pre-computed candidate list simultaneously, avoiding the need to update the list for each block and enabling parallel execution.
2Measurement precision
If the historical motion information candidate list is updated for each image block, then the prediction accuracy is improved, but the device complexity increases due to dependencies
Solution Approach 1:
By dividing the frame into independent image blocks and using a shared historical motion information candidate list, the patent eliminates complex inter-block dependencies. Each block can be processed independently with the same candidate list, significantly reducing the overall system complexity while maintaining prediction accuracy.
Solution Approach 2:
The same historical motion information candidate list serves all image blocks during prediction, making the list a universal resource. This multi-functional use of a single candidate list across multiple blocks reduces complexity by eliminating the need for separate candidate lists for each block.
3Productivity
If the historical motion information candidate list is not updated during prediction, then parallel processing efficiency is improved, but the prediction accuracy may deteriorate
Solution Approach 1:
The historical motion information candidate list is prepared in advance and maintained consistently throughout the prediction process. This preliminary preparation ensures that all image blocks can use the same accurate candidate list simultaneously, achieving both high prediction accuracy and parallel processing efficiency without the need for updates during prediction.
Data Source
AI summary
A bitstreams generating method includes obtaining a historical motion information candidate list used for encoding each of a plurality of image blocks included in a region of a current frame, encoding the plurality of image blocks according to the historical motion information candidate list, and generating bitstreams including one or more indexes of motion information. The historical motion information candidate list is a history-based motion vector prediction (HMVP) candidate list. During a process of performing prediction for all the plurality of image blocks included in the region, the historical motion information candidate list remains unchanged. Only the motion information of the last image block in the region is used to update the historical motion information candidate list after an encoding process of the last image block is finished.