Motion-Compensated Image Coding With Adaptive Virtual Sample Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image coding methods face inefficiencies in memory bandwidth and coding performance when dividing frames into small regions for motion compensated prediction, particularly due to increased memory requirements for calculating virtual samples at block boundaries.
Innovation Solution
The implementation of a moving image coding apparatus that adjusts the accuracy of virtual pixels based on region shape for motion compensated prediction, allowing for adaptive motion vector coding and reduced memory bandwidth by using a frame memory to store reference pictures and multiplexing shape information and motion vectors into the coded bitstream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the frame is divided into small regions for motion compensated prediction, then the prediction accuracy is improved, but the memory bandwidth requirement increases
Solution Approach 1:
The frame is divided into small regions (e.g., 8x8 pixel blocks) for motion compensated prediction, allowing independent motion estimation for each region. This segmentation improves prediction accuracy by capturing local motion variations while the invention manages the associated memory bandwidth increase through efficient reference picture storage and selective access patterns.
Solution Approach 2:
Different accuracy levels for virtual sample calculation are applied to different regions based on their shape and size. Small regions use half-pixel accuracy while larger regions can use quarter-pixel accuracy, optimizing the balance between prediction quality and memory bandwidth consumption across different parts of the frame.
2Measurement precision
If quarter-pixel accuracy virtual samples are calculated using 8-tap filters, then the prediction accuracy is improved, but the memory bandwidth increases significantly
Solution Approach 1:
The accuracy of virtual sample calculation is made dynamic and adaptive rather than uniform. The system automatically adjusts the filter tap count and accuracy level based on region characteristics, using 8-tap filters for quarter-pixel accuracy only when necessary (larger regions) and 4-tap filters for half-pixel accuracy for smaller regions, thereby optimizing memory bandwidth usage.
Solution Approach 2:
The invention changes the parameters of virtual sample calculation (filter tap count, accuracy level) based on region shape and size. By dynamically adjusting these parameters, the system achieves high prediction accuracy where needed while reducing memory bandwidth consumption in regions where lower accuracy suffices.
3Productivity
If high accuracy motion vectors are used, then the prediction efficiency is improved, but the code amount for motion vectors increases
Solution Approach 1:
The accuracy parameter of motion vectors is changed adaptively based on region characteristics. Small regions use half-pixel accuracy motion vectors while larger regions can use quarter-pixel accuracy, optimizing the balance between prediction efficiency and code amount by matching vector precision to regional requirements.
Data Source
AI summary
Each of an image coding apparatus and an image decoding apparatus uses a motion compensated prediction using virtual samples so as to detect a motion vector for each of regions of each frame of an input signal. Accuracy of virtual samples is locally determined while the accuracy of virtual samples is associated with the size of each region which is a motion vector detection unit in which a motion vector is detected. Virtual samples having half-pixel accuracy are used for motion vector detection unit regions having a smaller size 8×8 MC, such as blocks of 8×4 size, blocks of 4×8 size, and blocks of 4×4 size, and virtual samples having ¼-pixel accuracy are used for motion vector detection unit regions that are equal to or larger than 8×8 MC in size.