Motion Information Predictor Derivation for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video streams are fragile to transmission errors due to prediction methods, leading to bitstream de-synchronization issues when motion vector predictors are lost, causing decoding failures until the next re-synchronization frame.
Innovation Solution
A method to determine the number of motion information predictors based on data from the current slice rather than the motion vector value, eliminating duplicates and reducing parsing complexity, allowing for robustness to transmission errors and efficient bitstream decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion vector predictors are used for encoding, then encoding efficiency is improved, but transmission errors cause bitstream de-synchronization and decoding failures
Solution Approach 1:
The patent changes the parameter basis for determining the number of motion information predictors from motion vector values to slice-based data. This parameter change makes the predictor count independent of specific motion vector values, allowing the decoder to correctly parse the bitstream even when motion vector predictors are lost during transmission, thus resolving the de-synchronization issue while maintaining encoding efficiency
Solution Approach 2:
The patent uses a simplified, disposable approach by determining the number of motion information predictors based on easily accessible slice data rather than complex motion vector analysis. This lightweight method reduces parsing complexity and ensures that even if transmission errors occur, the decoder can quickly recover without requiring complex error correction mechanisms
2Measurement precision
If the number of motion information predictors is determined based on motion vector value, then prediction accuracy is improved, but parsing complexity increases
Solution Approach 1:
The patent extracts the determination of motion information predictor count from the motion vector value analysis process and relocates it to slice-based data. This separation simplifies the parsing process by removing the complex motion vector dependency while retaining sufficient accuracy for effective motion compensation
Solution Approach 2:
The patent segments the bitstream processing into slice-based units, determining the number of motion information predictors independently for each slice based on slice-specific data. This segmentation reduces overall parsing complexity by breaking down the global prediction problem into manageable local units
Data Source
AI summary
A method and device for deriving motion information of an image portion of an image predicted by motion prediction with respect to at least one reference image, the method comprising for said image portion: obtaining a number i of motion information predictors 808 of a second set of motion information predictors 814 taken from a first set of motion information predictors for decoding the image portion; generating said second set of motion information predictors 814 consisting of said number i of motion information predictors, wherein there are no duplicates among the i motion information predictors of said second set. A method and device for of decoding a bitstream, and a method and device for encoding a sequence of image including a method for deriving motion information of an image portion of an image predicted by motion prediction with respect to at least one reference image as described above.