Non-rectangular Block Partitioning for Video Decoding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies, such as H.264/AVC and HEVC, face inefficiencies in encoding and decoding due to increasing video image sizes, resolutions, and frame rates, requiring improved compression techniques for better encoding efficiency and image quality.
Innovation Solution
The method involves partitioning current blocks into non-rectangular blocks for inter and intra prediction, simplifying adaptive filtering by integrating sample adaptive offset and adaptive loop filtering, and determining inter and intra blocks based on a 1-bit flag to improve bit efficiency and remove discontinuity at block edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional video coding techniques (H.264/AVC, H.265/HEVC) are used, then coding efficiency is improved, but they cannot adapt to new flexible block partitioning structures
Solution Approach 1:
The patent extends the conventional AMBLE (Asymmetric Motion Boundary Extension) mode to handle diverse block partitioning structures including triangular, quincunx, and arbitrary quadrilateral partitions. The generalized AMBLE approach can predict blocks of various shapes and sizes by adapting the motion compensation process to different geometric configurations, making the coding technique universally applicable to both traditional and advanced partitioning schemes.
Solution Approach 2:
The patent introduces dynamic adaptation of prediction parameters based on the specific block partitioning structure being used. Rather than using fixed prediction modes, the system dynamically selects and adjusts motion compensation parameters according to the actual block geometry, allowing the coding technique to efficiently handle varying block shapes and sizes while maintaining coding performance.
2Measurement precision
If new flexible block partitioning structures are introduced, then prediction accuracy is improved, but decoding complexity increases due to lack of standardized techniques
Solution Approach 1:
The patent segments the decoding process into distinct stages: first identifying the block partitioning structure type, then selecting the appropriate prediction mode, and finally performing motion compensation with structure-specific parameters. This segmentation allows the decoder to handle complex partitioning structures through a systematic, step-by-step approach rather than requiring complex unified processing logic.
Solution Approach 2:
The patent introduces an intermediary classification mechanism that identifies the specific block partitioning structure type before applying the corresponding prediction technique. This intermediary step acts as a mediator between the diverse block structures and the prediction algorithms, routing each block type to its appropriate processing path and thereby simplifying the overall decoding complexity.
3Adaptability or versatility
If block partitioning into various shapes is implemented, then coding flexibility is improved, but existing prediction techniques become inapplicable
Solution Approach 1:
The patent changes the parameters of the motion compensation process to accommodate different block shapes. By adjusting prediction mode selection criteria, motion vector precision, and compensation window dimensions based on block geometry, the system maintains implementation simplicity while achieving coding flexibility across various partitioning structures.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A method and a device for inversely transforming transform coefficients of a current block are disclosed. A method for predicting a current block according to a first mode includes the steps of: on the basis of a partition mode syntax element, partitioning the current block into non-rectangular blocks; determining an intra-predicted intra block and an inter-predicted inter block among the non-rectangular blocks; and deriving prediction samples of a first area including the inter block on the basis of motion information, and deriving prediction samples of a second area including the intra block on the basis of an intra prediction mode.