Neural-Network Post-Filter SEI Signaling for Picture Rate Upsampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards lack efficient methods for signaling neural-network post-filter purposes, particularly in the context of picture rate upsampling, which affects the processing and transmission of digital video data.
Innovation Solution
The implementation of a neural-network post-filter (NNPF) with enhanced signaling through supplemental enhancement information (SEI) messages, allowing for more efficient conversion and processing of video data by specifying neural-network post-filter characteristics, including picture rate upsampling, in video coding systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If neural-network post-filter characteristics are signaled using traditional methods, then the video coding system can process video data, but the signaling efficiency and processing speed are insufficient
Solution Approach 1:
The patent changes the parameter representation by introducing a new SEI message type (NNPFC) with specific purpose codes (0-5) to indicate different neural-network post-filter operations. This parameter change enables efficient signaling of picture rate upsampling and other NNPF purposes, reducing signaling time while maintaining processing efficiency.
Solution Approach 2:
The patent uses SEI messages as an intermediary mechanism to carry neural-network post-filter characteristics between the encoder and decoder. This intermediary approach allows efficient communication of filtering parameters, including picture rate upsampling information, without direct complex signaling, thereby improving productivity while reducing time loss.
2Manufacturing precision
If picture rate upsampling is implemented with neural-network post-filters, then video quality improves, but the complexity of the coding system increases
Solution Approach 1:
The patent segments the neural-network post-filter processing into distinct purposes (visual quality improvement, chroma upsampling, resolution upsampling, picture rate upsampling, and combination modes). Each purpose is signaled separately through the SEI message, allowing the system to manage complexity by processing one purpose at a time while maintaining high video quality.
Solution Approach 2:
The patent creates a universal SEI message structure (NNPFC) that can signal multiple types of neural-network post-filter operations through a single message type with different purpose codes. This multi-functionality approach improves video quality through various filtering techniques while avoiding the need for multiple separate signaling mechanisms, thus managing system complexity.
3Adaptability or versatility
If multiple types of upsampling are combined, then more flexible video processing is achieved, but the signaling and processing becomes more complex
Solution Approach 1:
The patent implements dynamic signaling where the SEI message includes a purpose code that can indicate different combinations of upsampling operations (chroma upsampling, resolution upsampling, picture rate upsampling). This dynamic approach allows flexible video processing adapted to different scenarios while using a standardized signaling mechanism that manages processing complexity through clear purpose identification.
Data Source
AI summary
A mechanism for processing video data is disclosed. The mechanism includes determining a neural-network post-filter (NNPF) purpose based on a neural-network post-filter characteristics (NNPFC) supplemental enhancement information (SEI) message. A conversion is performed between a visual media data and a bitstream based on the NNPF purpose. Multiple input pictures are used for the NNPF purpose, and the NNPF is enabled to selectively generate output pictures for some input picture(s) and not to generate output pictures for other input picture(s).