Neural-Network Post-Filter Signaling for Reliable Picture Output
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images leads to a significant increase in transmission and storage costs due to the rise in the amount of transmitted information, necessitating high-efficient image compression technology.
Innovation Solution
An image encoding/decoding method that utilizes neural-network post-filter characteristics (NNPFC) SEI messages to determine whether output pictures are generated and encoded, along with neural-network post-filter activation (NNPFA) SEI messages to specify output conditions, improving encoding/decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images are transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent extracts and transmits only the essential visual information needed to reconstruct high-quality images at the receiver end, rather than transmitting complete high-resolution image data. This is achieved through selective transmission of parameters and use of generative models that can reconstruct full-quality images from compressed representations, thereby reducing transmission volume while maintaining image quality.
Solution Approach 2:
The patent performs preliminary processing of image data at the transmitter end to extract key features and parameters before transmission. By pre-processing the image data to identify and transmit only the most important visual information, the system reduces the amount of data that needs to be transmitted while ensuring that sufficient information is available to reconstruct high-quality images at the receiver end.
2Adaptability or versatility
If NNPFA is activated to generate output pictures, then image processing capability is improved, but picture output may fail under certain conditions
Solution Approach 1:
The patent implements dynamic control of the NNPFA activation based on the purpose of the neural network post-filter and the number of input pictures. The system adaptively determines whether to activate NNPFA and whether to output pictures in real-time, adjusting its behavior according to the specific processing scenario. This dynamic approach ensures that picture output is reliable and appropriate for each specific case, preventing failures that would occur with static activation rules.
Solution Approach 2:
The patent incorporates feedback mechanisms that monitor the processing state and output conditions to determine whether pictures should be output. By continuously assessing the purpose of the NNPF and the number of input pictures, the system provides feedback control over NNPFA activation and picture output, ensuring that processing capability is optimized while maintaining reliable picture output in all scenarios.
Data Source
AI summary
An image encoding/decoding method, a method of transmitting a bitstream and a computer-readable recording medium storing a bitstream are provided. The image encoding method may comprise encoding first information indicating whether at least one output picture corresponding to at least one input picture is generated from an NNPFC (neural-network post-filter characteristics) SEI (supplemental enhancement information) message and encoding second information indicating whether the generated output picture is output from an NNPFA (neural-network post-filter activation) SEI message. A value of the second information may be determined based on at least one of whether a purpose of an NNPF (neural-network post-filter) is picture rate upsampling or whether the number of input pictures is multiple.