Boundary Patch Processing for Spatial Extrapolation in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in efficiently calculating patch sizes for spatial extrapolation, particularly in advanced video coding systems like ITU-T H.266, which can impact the quality and efficiency of video decoding processes.
Innovation Solution
The proposed techniques involve parsing neural-network post-filter characteristics messages to determine spatial extrapolation offsets and adjust patch sizes based on input picture heights and chroma formats, ensuring accurate calculation and modification of output patch heights to enhance decoding performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If spatial extrapolation is used to generate content outside the spatial area of input pictures, then the quality of reconstructed video data is improved, but the complexity of the decoding process increases due to patch size calculation and modification requirements
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting patch sizes based on spatial extrapolation offsets and picture dimensions. The output patch height is calculated and modified using formulas that incorporate input picture height, offset values, and chroma format parameters, allowing the system to adapt patch parameters to different decoding scenarios and maintain optimal reconstruction quality without fixed rigid structures
Solution Approach 2:
The patent segments the video decoding process into distinct stages: parsing neural-network post-filter characteristics messages, determining spatial extrapolation offsets, calculating initial output patch heights, and applying conditional modifications. This segmentation allows each stage to be processed independently with specific optimization, reducing overall computational complexity while maintaining quality
2Measurement precision
If patch sizes are dynamically adjusted based on spatial extrapolation offsets and chroma formats, then the accuracy of video decoding is improved, but the computational time required increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating output patch heights using standardized formulas before actual video reconstruction. The system determines spatial extrapolation offsets and computes initial patch dimensions in advance, preparing all necessary parameters beforehand to avoid complex real-time calculations during the main decoding process, thus reducing computational time while maintaining accuracy
Solution Approach 2:
The system efficiently manages parameter changes by using conditional logic that only modifies output patch heights when specific conditions are met (when top spatial extrapolation offset is less than 0 and vertical sample coordinate is less than or equal to the negative of the offset). This selective parameter modification reduces unnecessary computations while maintaining decoding accuracy where needed
Data Source
AI summary
A device may be configured to perform spatial extrapolation based on information included in a neural-network post-filter characteristics message. In one example, a neural-network post-filter characteristics message includes a purpose syntax element indicating a purpose of a neural-network post-filter includes spatial extrapolation. The device may be configured to calculate a output patch height and a output patch width based on a purpose of the neural-network post-filter including spatial extrapolation. In a case where an input patch is on or near a border, a calculated output patch size is adjusted based on a spatial extrapolation offset.