ISP Sub-partition Planar Intra Prediction for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently encoding and decoding video data, particularly in reducing redundancy and improving compression ratios, especially with the increasing complexity of intra prediction modes and motion compensation techniques.
Innovation Solution
The method involves partitioning a current block into sub-partitions based on an intra sub-partition mode, using different or shared reference samples for reconstruction, and employing planar or DC intra prediction modes to optimize video encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional intra coding without intra prediction is used, then the coding process is simpler, but the compression ratio is lower and more bits are required to represent the block
Solution Approach 1:
The current block is divided into multiple sub-partitions (e.g., 2×2 sub-partitions for a 4×4 block), and each sub-partition is independently predicted using intra prediction techniques. This segmentation allows the patent to apply prediction modes to smaller regions, improving compression efficiency while maintaining manageable complexity through localized processing.
Solution Approach 2:
The patent performs intra prediction on sub-partitions before the main transformation and coding stages. By pre-calculating prediction values for each sub-partition using reference samples from already decoded neighboring blocks, the method reduces the amount of data that needs to be encoded, thereby improving compression ratio without significantly increasing overall complexity.
2Loss of information
If intra prediction techniques using surrounding sample data are employed, then the compression ratio is improved, but the device complexity and processing requirements increase
Solution Approach 1:
Different prediction modes (e.g., planar, DC, angular) are applied to different sub-partitions based on their local characteristics. Each sub-partition is independently processed with the most appropriate prediction mode, allowing the system to optimize compression for each region while keeping the overall complexity manageable through localized decision-making.
Solution Approach 2:
The patent applies intra prediction to only certain sub-partitions rather than uniformly to the entire block. By selectively applying prediction techniques to sub-partitions where they provide the most benefit, the method achieves improved compression ratios while avoiding the full processing overhead of applying prediction to every possible region.
3Measurement precision
If the current block is partitioned into multiple sub-partitions with different reference samples, then the prediction accuracy is improved, but the data requirements and processing complexity increase
Solution Approach 1:
The block is segmented into sub-partitions, each using its own reference samples from neighboring already-decoded blocks. This segmentation enables each sub-partition to be predicted with higher accuracy using locally relevant reference data, while the total data requirement is managed by reusing reference samples across multiple sub-partitions where applicable.
Solution Approach 2:
The patent changes the prediction parameters (reference sample selection, prediction mode) for each sub-partition based on local characteristics. By adapting the prediction parameters to match the local content of each sub-partition, the method achieves higher prediction accuracy without proportionally increasing data requirements, as the same reference samples can serve multiple sub-partitions.
Data Source
AI summary
Aspects of the disclosure provide methods, apparatuses, and non-transitory computer-readable storage mediums for video encoding/decoding. Prediction information for a current block in a current picture is determined. The prediction information indicates that the current block is to be encoded in (1) an intra sub-partition (ISP) mode and (2) one of a planar intra prediction mode or a DC intra prediction mode. The current block is partitioned into a plurality of sub-partitions based on the ISP mode. Each sub-partition is associated with at least one different reference sample. Residual information for each sub-partition is generated based on the at least one different reference sample associated with the respective sub-partition. The current block is encoded based on the determined prediction information and the generated residual information of the plurality of sub-partitions.