Neural-Network Post-Filter Signaling for Clear Output Picture Decoding
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 technologies.
Innovation Solution
An image encoding/decoding method that includes obtaining and signaling post-filter-based output picture information through neural-network post-filter characteristics (NNPFC) SEI messages, with specific conditions for output picture generation, to improve encoding/decoding efficiency and clarify the meaning of information related to the output picture.
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 difference information (residual signals) between original and reconstructed images rather than transmitting complete high-resolution image data. This allows high-quality reconstruction with reduced transmission volume by separating and transmitting only the necessary corrective information.
Solution Approach 2:
The patent transforms image data into different parameter representations (frequency domain coefficients, residual differences) that enable more efficient compression. By changing the representation parameters from spatial domain pixel values to transformed domain coefficients, the patent achieves better compression ratios while maintaining image quality.
2Manufacturing precision
If neural-network post-filter is applied to improve coding quality, then coding quality is improved, but decoder error and ambiguity in output picture specification increase
Solution Approach 1:
The patent performs preliminary actions by explicitly specifying all necessary parameters for neural-network post-filter operation in advance through SEI messages. The output picture specification, including picture index, resolution, and format information, is predetermined and communicated to the decoder before reconstruction, eliminating ambiguity and ensuring reliable reproduction.
Solution Approach 2:
The patent implements feedback mechanisms where the encoder provides detailed specification information about the neural-network post-filter output pictures through SEI messages. This feedback loop ensures the decoder can accurately reproduce the intended output by receiving explicit guidance on picture parameters, resolution, and format, thereby reducing decoder errors.
3Loss of information
If detailed output picture information is signaled in SEI messages, then output picture specification clarity is improved, but message complexity and processing overhead increase
Solution Approach 1:
The patent segments the output picture specification information into distinct, organized fields within SEI messages. Each aspect (picture index, resolution, format) is separated into specific syntax elements, making the information structured and easier to process. This segmentation reduces processing complexity by providing clear, discrete data elements rather than unstructured detailed information.
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. An image decoding method may comprise obtaining post-filter-based corresponding output picture information for an input picture from an NNPFC (neural-network post-filter characteristics) SEI (supplemental enhancement information) message and obtaining the corresponding output picture for the input picture based on the corresponding output picture information. The corresponding output picture information may include output picture generation information about whether the corresponding output picture of a post-filter for the input picture is generated.