Asymmetric Discrete Sine Transform for Video Intra-Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 2D-DCT is inefficient for video data encoding as it maximizes at both ends and is agnostic to the statistical characteristics of prediction residuals, limiting the effectiveness of template-matched intra-coding schemes.
Innovation Solution
The use of an asymmetric discrete sine transform (ADST) for encoding and decoding video data, which selects a matched template based on a weighted scheme of reconstructed adjacent pixels to generate residuals, optimizing coding performance by considering the statistical characteristics of the residuals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If 2D-DCT is used for encoding prediction residuals, then the encoding process is simple and standardized, but the coding efficiency is poor because DCT basis functions are agnostic to statistical characteristics of residuals
Solution Approach 1:
The patent changes the transform parameters by switching from symmetric DCT basis functions to asymmetric ADST basis functions that are specifically designed to match the statistical characteristics of prediction residuals. This parameter change allows the transform to adapt to the actual data distribution, improving coding efficiency while maintaining computational feasibility through established ADST algorithms.
Solution Approach 2:
The patent applies asymmetry by using asymmetric discrete sine transform (ADST) basis functions instead of the symmetric DCT basis functions. The ADST basis functions are designed to be asymmetric to match the asymmetric statistical characteristics of prediction residuals in video coding, where residuals typically have different distributions at different positions within the block.
2Measurement precision
If template matching is used to select reference blocks, then prediction accuracy is improved, but the computational complexity increases due to the weighted scheme of reconstructed adjacent pixels
Solution Approach 1:
The patent applies local quality by using a weighted scheme in template matching where different pixels in the reference block are assigned different weights based on their proximity to the current block. Pixels closer to the current block receive higher weights, allowing the matching process to focus on the most relevant areas and improve prediction accuracy while managing computational complexity through selective weighting.
Data Source
AI summary
A system includes an encoder and a decoder. The encoder selects a first matched template for un-encoded pixels of a video frame using an algorithm for measuring a similarity between image blocks of the video frame, the algorithm being based on a weighted scheme of reconstructed adjacent pixels, generates at least one residual for the un-encoded pixels of the video frame based on the matched template, and encodes residuals as compressed bits using an asymmetric discrete sine transform (ADST). The decoder decodes the compressed video bits as residuals using the ADST, selects a second matched template using an algorithm for measuring a similarity between image blocks of a video frame associated with the compressed video bits, the algorithm being based on a weighted scheme of reconstructed adjacent pixels, and generates reconstructed pixels of the video frame based on the matched template and the decoded compressed video bits.