Motion Vector Prediction Scaling for Multi-Reference Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In conventional moving picture encoding systems, the use of median motion vectors from surrounding coded regions with different reference frames leads to significant differences between predicted and actual motion vectors, resulting in decreased efficiency of information compression and decoding.
Innovation Solution
A moving picture encoding and decoding apparatus that divides frame images into target regions, detects motion vectors using multiple reference frames, and corrects and scales these vectors based on temporal relations and time information to predict an optimum motion vector, reducing the difference between predicted and actual motion vectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If median motion vectors from surrounding coded regions with different reference frames are used for prediction, then the prediction process is simple, but the accuracy of motion vector prediction deteriorates significantly
Solution Approach 1:
The patent changes the parameters of motion vector prediction by introducing temporal relation analysis and scaling factors based on time differences between reference frames. Instead of directly using median motion vectors, the system scales motion vectors according to temporal relationships (time differences between reference frames and current frame) and selects the most appropriate predicted motion vector based on these scaled values, thereby improving prediction accuracy while maintaining reasonable complexity
Solution Approach 2:
The patent segments the motion vector prediction process into distinct stages: (1) collecting motion vectors from surrounding regions, (2) analyzing temporal relations between reference frames, (3) scaling motion vectors based on time differences, (4) selecting optimal predicted motion vectors. This segmentation allows each stage to be optimized independently, improving overall accuracy without proportionally increasing complexity
2Productivity
If motion vectors from surrounding regions are used without temporal correction, then the processing speed is fast, but the efficiency of information compression deteriorates
Solution Approach 1:
The patent performs preliminary temporal relation analysis and scaling factor calculation before final motion vector prediction. By pre-computing time differences between reference frames and determining scaling factors in advance, the system prepares corrected motion vectors that will be more accurate during the actual encoding process, improving compression efficiency without significantly impacting processing speed
Solution Approach 2:
The patent replaces the simple mechanical averaging process (median calculation) with a more sophisticated temporal correction mechanism that uses scaling factors based on time differences. This substitution transforms the prediction process from a purely spatial operation to one that incorporates temporal dynamics, improving compression efficiency by better accounting for motion changes over time
Data Source
AI summary
An object of the present invention is to increase efficiency of information compression in coding and decoding. A moving picture encoding apparatus 10 of the present invention has a motion vector predicting part for performing, based on a temporal relation among adjacent reference frame images 703a, 703b, 703c referred to for detecting motion vectors of adjacent blocks adjacent to a coding target block, a target reference frame image 702 referred to for detecting a motion vector of the target block, and a target frame image 701 being the frame image of the coding target, or based on time information thereof, a correction of scaling the motion vectors 751a, 751b, 751c of the adjacent blocks on the basis of the target reference frame image 702; and a determination of an optimum predicted motion vector based on the motion vectors of the adjacent blocks, and thereby predicting the optimum predicted motion vector after the correction.