Video Coding With Cross-Component Filtering and Direction Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in representing intra prediction directions, particularly with increased complexity and bit usage for less likely directions, and motion vector prediction mechanisms that do not fully leverage statistical likelihoods in video content.
Innovation Solution
Implement advanced intra prediction techniques with up to 65 directions and improved motion vector prediction methods, such as spatial merge, to reduce redundancy and bit usage by predicting intra and inter-block relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional intra prediction techniques are used, then device complexity is reduced, but bit usage efficiency deteriorates due to inability to represent all intra prediction directions effectively
Solution Approach 1:
The patent changes the parameter of prediction direction count from traditional limited directions to up to 65 directions, allowing more precise representation of intra prediction directions. This parameter change enables better bit usage efficiency by matching the actual directional characteristics of video content while managing complexity through selective application.
Solution Approach 2:
The patent introduces dynamic selection of prediction directions based on statistical likelihoods observed in video content. The system adaptively chooses between different numbers of prediction directions (including up to 65) depending on the specific coding context, allowing the complexity to vary dynamically rather than being fixed, thus resolving the contradiction between information loss and device complexity.
2Productivity
If motion vector prediction mechanisms are simplified, then device complexity is reduced, but compression efficiency deteriorates due to inability to leverage statistical likelihoods
Solution Approach 1:
The patent implements feedback mechanisms where the system observes statistical likelihoods of motion vectors from previously decoded blocks and uses this feedback to inform prediction decisions for current blocks. This feedback loop enables the system to adaptively improve compression efficiency by leveraging observed patterns without requiring overly complex predefined rules.
Solution Approach 2:
The motion vector prediction mechanism uses information from the video content itself (statistical likelihoods observed during decoding) to improve its own performance. The system serves itself by using its own decoded output as input for better prediction, eliminating the need for external complex control mechanisms while improving compression efficiency.
3Measurement precision
If more intra prediction directions are supported, then measurement precision is improved, but bit usage increases for less likely directions
Solution Approach 1:
The patent applies local quality by providing high prediction direction precision (up to 65 directions) only where needed based on statistical likelihoods. Rather than uniformly supporting all 65 directions for all blocks, the system selectively applies higher precision to regions where the statistics indicate benefit, while using fewer directions elsewhere, thus optimizing the balance between measurement precision and bit usage.
Solution Approach 2:
The patent implements partial action by supporting up to 65 prediction directions but not necessarily all of them for every coding block. The system applies the full range of directions selectively based on statistical likelihoods, using the complete set only when the content characteristics justify the additional bit usage, and using fewer directions when they suffice, thus avoiding excessive bit consumption while maintaining precision when needed.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
Aspects of the disclosure include methods, apparatuses, and non-transitory computer-readable storage mediums for video encoding/decoding. An apparatus includes processing circuitry decoding prediction information of a current block, which indicates a plurality of combinations of processed reconstructed samples and corresponding output values of a filter process. The processing circuitry performs the filter process on processed reconstructed samples of a first color component of the current block. The processed reconstructed samples of the first color are not included in the plurality of combinations of the processed reconstructed samples. The processing circuitry determines an output value of the filter process based on one of a default value, a predicted output value, and one of the output values indicated in the prediction information. The processing circuitry reconstructs the current block based on the output value being applied to a second color component of the current block that is different from the first color component.