Neural-Network Post-Filter SEI Signaling for Video Downsampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards lack support for neural-network post-filter purposes with downsampling capabilities, particularly in combination with upsampling or downsampling of output picture size and chroma format, which can enhance video processing performance.
Innovation Solution
The method involves defining new neural-network post-filter purposes for chroma format and output picture size downsampling or upsampling, including specific syntax elements to signal these operations, such as absolute increases or decreases in picture dimensions and chroma format changes, using SEI messages in video bitstreams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If neural-network post-filter purposes with downsampling capabilities are added, then video processing performance is improved, but device complexity increases
Solution Approach 1:
The patent extends the existing nnpfc_purpose parameter by adding new purpose values (5-9) that enable downsampling operations. This allows the neural-network post-filter to handle additional processing scenarios including output picture size downsampling and chroma format downsampling, thereby improving video processing performance while maintaining compatibility with existing standards through parameter extension rather than structural complexity increase
2Adaptability or versatility
If new nnpfc_purpose values for downsampling are defined, then adaptability is improved, but standard complexity increases
Solution Approach 1:
The patent implements multi-functionality by allowing a single extended parameter (nnpfcPurpose) to represent multiple downsampling scenarios through different purpose values. This includes output picture size downsampling (purposes 5-7) and chroma format downsampling (purposes 8-9), enabling the standard to handle diverse downsampling needs through a unified signaling mechanism without requiring separate complex structures for each function
Data Source
AI summary
A method for processing media data is disclosed. In an embodiment, the method includes determining a value of a neural-network post-filter characteristics (NNPFC) purpose (nnpfc_purpose) in a NNPFC supplemental enhancement information (SEI) message, wherein the nnpfc_purpose is configured to be set to include output picture size downsampling, and wherein at least one of (1) an output picture width is not equal to an input picture width and (2) an output picture height is not equal to an input picture height; and performing a conversion between a visual media data and a bitstream based on the nnpfc_purpose.