L-Shaped Block Transforms for Lower-Complexity Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like VVC and HEVC face inefficiencies in compression due to sub-optimal partitioning of blocks, particularly when using L-shaped partitions, which do not allow for straightforward application of traditional orthogonal transforms, leading to increased computational complexity and sub-optimal compression efficiency.
Innovation Solution
Introduce a method for partitioning image blocks into at least two partitions, with one being an L-shaped partition, and design transforms specifically tailored for these configurations to improve compression efficiency by reducing computational complexity and enhancing encoding/decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional orthogonal transforms are applied to L-shaped partitions, then compression efficiency is improved, but computational complexity increases significantly
Solution Approach 1:
The L-shaped partition is segmented into multiple rectangular sub-partitions (e.g., two or more rectangles) that can be independently transformed using standard orthogonal transforms. This segmentation allows the application of efficient rectangular transform algorithms to the irregular L-shaped region without requiring complex custom transform implementations, thereby maintaining compression efficiency while reducing computational complexity.
Solution Approach 2:
Multiple rectangular sub-partitions are merged to collectively represent the L-shaped partition. By processing several standard rectangular regions together as a unified L-shaped block, the method achieves compression efficiency comparable to dedicated L-shaped transforms while leveraging the computational efficiency of standard rectangular transform algorithms.
2Productivity
If L-shaped partitions are used for block splitting, then rate-distortion performance is improved, but ease of operation deteriorates due to lack of straightforward transform application
Solution Approach 1:
The L-shaped partition is divided into multiple rectangular sub-partitions that can be independently processed using standard transform algorithms. This segmentation maintains the rate-distortion benefits of L-shaped partitioning while restoring ease of operation by enabling the use of well-established, straightforward rectangular transform implementations.
Solution Approach 2:
The method enables universal application of standard rectangular transform algorithms to L-shaped partitions by treating them as collections of rectangular sub-partitions. This multi-functional approach allows the same transform code to handle both regular rectangular blocks and L-shaped partitions, significantly improving ease of operation while preserving rate-distortion performance.
3Productivity
If custom transforms are designed for L-shaped partitions, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
Instead of designing a custom transform for the entire L-shaped partition, the method segments the L-shape into rectangular sub-partitions and applies standard transforms to each segment. This approach achieves compression efficiency similar to custom L-shaped transforms while avoiding the implementation complexity of designing and coding new transform algorithms.
Solution Approach 2:
The method copies and reuses existing standard rectangular transform implementations rather than creating new custom transforms for L-shaped partitions. By applying proven rectangular transform algorithms to segmented L-shaped regions, the method achieves comparable compression efficiency without the device complexity of implementing novel transform mathematics.
Data Source
AI summary
An encoding method (decoding method respectively) is disclosed wherein an image block to be encoded (decoded respectively) is partitioned in at least two partitions, at least one of said partition has an L-shape. Various configurations are defined based on the location of the L-shape in the image block. To reduce the computation complexity, only a subset of the configurations may be allowed. A transform and an inverse transform are designed to be applied on such L-shaped partition for encoding and decoding.