Intra-Block Copy Prediction for Camera-Captured Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The current design of the intra-block-copy (IBC) prediction mode in video coding standards like VVC is more adapted to graphical and screen content, limiting the compression performance improvement when applied to camera-captured video contents.
Innovation Solution
Adapt the IBC prediction mode to better suit camera-captured video content by configuring it to predict blocks based on reference blocks within the video picture, allowing for improved compression efficiency on both screen and natural video content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the IBC prediction mode is designed for graphical and screen content, then coding efficiency for screen content is improved, but compression performance for camera-captured video content deteriorates
Solution Approach 1:
The patent makes the IBC prediction mode dynamic by introducing a mode type parameter that can switch between screen content configuration and camera-captured video content configuration. The system adapts its prediction behavior based on the content type, allowing optimal performance for both graphical/screen content and natural video content through configurable prediction modes.
2Measurement precision
If block matching is performed at encoder to find optimal block vector for each block, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by differentiating the prediction approach based on content type. For screen content, one optimization strategy is used, while for camera-captured video content, a different configuration is applied. This allows the system to tailor the computational approach to the local characteristics of the video content, improving accuracy where needed while reducing unnecessary complexity.
Data Source
AI summary
A method of predicting a block of a video picture, includes: signaling an intra-block-copy mode type, the intra-block-copy mode type indicating whether an intra-block-copy prediction mode is configured to predict camera-captured video content or not, the intra-block-copy prediction mode determining at least one block vector for predicting the block of the video picture from at least one reference block of the video picture; configuring the intra-block-copy prediction mode to predict the block of the video picture, in response to the intra-block-copy mode type indicating that the intra-block-copy prediction mode is configured to predict the camera-captured video content; and deriving a predicted block of the block of the video picture based on the configured intra-block-copy mode prediction mode.