Geometric Intra Prediction Mode for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding methods have limitations in intra prediction efficiency due to the constraint of using limited pixels in the same picture for generating prediction signals, leading to relatively low performance compared to inter prediction techniques, especially with increasing image sizes, resolutions, and frame rates.
Innovation Solution
A video coding method and apparatus that apply a geometric intra-prediction mode by generating two intra predictors using different intra-prediction modes and weighting them based on pixel-wise weights derived from a geometric block partition to improve prediction performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional intra prediction techniques are used with limited pixels in the same picture, then the method is simple to implement, but the prediction performance is relatively low compared to inter prediction techniques
Solution Approach 1:
The current block is divided into multiple sub-blocks, and different intra prediction modes are applied to different sub-blocks. This segmentation allows the patent to overcome the limitation of using limited pixels by applying specialized prediction strategies to specific regions, thereby improving overall prediction performance without requiring additional reference pixels from other pictures.
Solution Approach 2:
Different intra prediction modes are selected and applied to different sub-blocks based on their local characteristics. This local quality approach enables each sub-block to use the most appropriate prediction method for its specific content, improving prediction accuracy while maintaining simplicity within each local region.
2Reliability
If multiple line buffers and MRL intra prediction techniques are used to improve prediction performance, then the prediction accuracy is enhanced, but the device complexity and computational load increase
Solution Approach 1:
The patent extracts and utilizes already-decoded neighboring pixels that are spatially adjacent to the current block, without requiring additional multiple line buffers or complex MRL mechanisms. By taking out and effectively using the available neighboring pixels through sub-block segmentation and mode selection, the patent improves prediction accuracy while avoiding the additional complexity of multiple buffer structures.
3Reliability
If the image size, resolution, and frame rate are increased to improve video quality, then the video quality is enhanced, but the amount of data to be encoded increases
Solution Approach 1:
By segmenting the current block into multiple sub-blocks and applying different intra prediction modes to each, the patent improves coding efficiency for high-resolution images. This allows more accurate representation of local variations in high-detail content, reducing the residual data that needs to be encoded and transmitted, thereby managing data quantity while maintaining enhanced video quality.
Solution Approach 2:
The patent changes the prediction parameters by selecting different intra prediction modes for different sub-blocks based on local characteristics. This parameter adaptation improves the accuracy of prediction for high-resolution content, reducing the encoding burden and managing data quantity requirements while maintaining high video quality.
Data Source
AI summary
A video coding method and a video coding apparatus are configured for video coding using a geometric intra-prediction mode. The video coding method and the video coding apparatus generate two intra predictors by using two different intra-prediction modes. The video coding method and the video coding apparatus use pixel-wise weights based on a geometric block partition to weight sum two intra predictors and thereby generate a final intra predictor.