Motion Information Refinement for Projection Format Distortion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding/decoding technologies face challenges in achieving enhanced compression efficiency, particularly in refining distortions generated by projection formats such as HEC, which affect the accuracy of motion information prediction and coding efficiency.
Innovation Solution
A method and apparatus for refining motion information distortion in video encoding/decoding, which involves decoding information for motion prediction, predicting motion information, refining the motion information using decoded and predicted data, and reconstructing blocks based on refined motion information. This process is signaled through sequence and picture parameter sets and can utilize pixel values from L0 and L1 reference pictures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion information prediction is performed using conventional methods, then decoding speed is maintained, but compression efficiency is insufficient due to distortion in projection formats
Solution Approach 1:
The patent applies preliminary action by performing motion information refinement before motion compensation. The decoder refines motion information by compensating for projection format distortions (such as HEC distortions) before using the motion information to predict the current block. This preliminary refinement improves compression efficiency by ensuring more accurate motion prediction, reducing the residual error that needs to be encoded.
Solution Approach 2:
The patent changes parameters by modifying motion information parameters to account for projection format distortions. Specifically, the decoder adjusts motion vector parameters by calculating distortion compensation values based on the projection format type and block position. This parameter adjustment allows the motion prediction to accurately reflect the distorted geometry of projection formats like HEC, thereby improving compression efficiency without requiring complex additional processing.
2Measurement precision
If motion information is refined using projection format distortion compensation, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies local quality by applying distortion compensation selectively based on local characteristics. The decoder determines whether refinement is needed by checking if the current block is subject to projection format distortion (e.g., based on block position and projection format type). Only blocks that require refinement undergo the additional processing, while other blocks use conventional motion prediction. This selective approach improves prediction accuracy where needed while minimizing overall processing complexity.
Solution Approach 2:
The patent uses copying by creating a compensated motion information value based on an original motion information value and a distortion compensation value. The decoder copies the basic motion prediction process and adds a correction layer, rather than completely redesigning the motion prediction mechanism. This approach maintains the simplicity of the original motion estimation while improving accuracy through additive distortion compensation.
3Productivity
If distortion refinement is performed for all blocks, then compression efficiency is maximized, but processing time increases
Solution Approach 1:
The patent applies partial action by performing distortion refinement only on a subset of blocks that benefit most from it. The decoder uses criteria such as block position, projection format type, and motion prediction mode to determine which blocks require refinement. By applying refinement partially rather than universally, the patent achieves significant compression efficiency improvements for the refined blocks while avoiding the excessive processing time that would result from refining all blocks.
Data Source
AI summary
There is provided a method of decoding an image, the method comprising: decoding information for motion information prediction of a current block from a bitstream, predicting motion information of the current block based on the information, refining motion information of the current block by using the decoded information and the predicted motion information of the current block and reconstructing the current block based on the refined motion information of the current block.