Distance Weighted Bi-Directional Intra Prediction for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as HEVC/H.265, face limitations in coding efficiency for intra prediction due to the reliance on a single adjacent row or column of pixels for prediction, which results in higher bitrates for achieving the same visual quality.
Innovation Solution
An apparatus and method for intra prediction that generates secondary reference pixel values based on primary reference pixel values from neighboring video coding blocks, allowing for more comprehensive prediction and improved coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single adjacent row or column of pixels is used for intra prediction, then the device complexity is reduced, but the coding efficiency deteriorates resulting in higher bitrates
Solution Approach 1:
The reference pixel values are segmented into two distinct groups: primary reference pixel values from directly adjacent blocks and secondary reference pixel values from non-adjacent blocks. This segmentation allows the patent to utilize multiple reference sources without creating a monolithic complex structure, thereby improving coding efficiency while maintaining manageable device complexity through organized modular processing
Solution Approach 2:
The patent implements a nested structure where secondary reference pixel values are generated based on primary reference pixel values through gradient calculations. The secondary references are nested within the primary reference framework, creating a hierarchical reference system that enhances prediction accuracy without proportionally increasing device complexity
2Loss of information
If more reference pixel values are used for prediction, then the coding efficiency is improved, but the computational complexity increases
Solution Approach 1:
The patent performs preliminary calculations of gradient values and secondary reference pixel values before the actual prediction process. By pre-computing these reference values and storing them for later use, the patent reduces the computational burden during the main prediction phase, thereby improving coding efficiency without proportionally increasing overall computational complexity
Solution Approach 2:
Different reference pixel values are used for different regions of the current block based on their spatial relationships. Primary reference pixels are used for regions closer to adjacent blocks, while secondary reference pixels are used for regions farther away. This local differentiation optimizes the prediction quality for each region without requiring all reference values to be processed equally, thus managing computational complexity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to an apparatus (100) for intra prediction of a current video coding block of a frame of a video signal, wherein the current video coding block comprises a plurality of pixels and each pixel is associated with a pixel value. The apparatus (100) comprises: a reference pixel unit (101) configured to generate on the basis of a plurality of primary reference pixel values a plurality of secondary reference pixel values, wherein the plurality of primary reference pixel values are associated with a plurality of primary reference pixels located in neighboring video coding blocks of the current video coding block, wherein the reference pixel unit is configured to generate each of the secondary reference pixel values on the basis of two or more of the primary reference pixel values; and an intra prediction unit (103) configured to intra predict the pixel values of the pixels of the current video coding block on the basis of the plurality of primary reference pixel values and the plurality of secondary reference pixel values.