Neural-Network Post-Filter Output Control in Image 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 or bits, necessitating high-efficient image compression technology.
Innovation Solution
An image encoding/decoding method that includes processing neural-network post-filter (NNPF) related supplemental enhancement information (SEI) messages to improve encoding/decoding efficiency, clarify output picture information, and reduce decoder errors by determining whether to output an image based on input picture presence, thereby optimizing the filtering process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality image data is transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent extracts and transmits only the residual difference between the original image and the predicted image, rather than transmitting the entire high-resolution image. This residual data contains only the essential information that cannot be predicted, significantly reducing the amount of transmitted data while maintaining image quality.
Solution Approach 2:
The patent replaces traditional compression mechanisms with a neural network-based prediction mechanism. The encoder uses a neural network to predict image content, and the transmitter sends only the prediction error, substituting mechanical compression algorithms with intelligent prediction algorithms that adapt to content characteristics.
2Measurement precision
If neural-network post-filter is applied to improve coding quality, then decoding accuracy is improved, but device complexity increases
Solution Approach 1:
The patent performs post-filtering operations using neural networks after the main decoding process. By applying the neural network filter as a preliminary or subsequent step rather than integrating it into the core decoding algorithm, the system improves decoding accuracy while keeping the main decoding process simple and efficient.
Solution Approach 2:
The neural network post-filter acts as an intermediary between the decoded image and the final output. It processes the decoded image to correct prediction errors and improve quality without requiring changes to the main decoding architecture, thus improving accuracy while maintaining structural simplicity.
Data Source
AI summary
Provided are an image encoding/decoding method, a bitstream transmission method, and a computer-readable recording medium storing a bitstream. The image decoding method according to the present disclosure comprises the steps of: acquiring post-filter-based output picture information for an input picture from a neural-network post-filter (NNPF)-related supplemental enhancement information (SEI) message; and on the basis of the output picture information, acquiring an output picture for the input picture, wherein the output picture information may include output picture output information which is information indicating whether or not the output picture for the input picture is output.