Decoded image enhancement method

Abstract

The invention discloses a decoded image enhancement method, and mainly solves the problem that the current decoded enhanced image algorithm does not fully consider the visual and transmission requirements of human eyes on different areas of an image and cannot achieve the balance of texture subjective quality and edge signal fidelity. The method is based on different characteristics of a texture region and an edge region. An edge region-based decoding enhancement method and a texture region-based decoding enhancement method are respectively designed to obtain a basic fidelity layer and a texture enhancement layer; and then the basic fidelity layer and the texture enhancement layer are adaptively fused by using a region adaptive fusion technology, so that the balance of texture subjective quality and edge signal fidelity is obtained, a decoded image is enhanced better, and the subjective experience of a user and the fidelity transmission of a signal are improved.

Description

technical field [0001] The invention relates to the technical field of image processing, and in particular to an image decoding enhancement system based on a deep neural network, which belongs to decoding image reconstruction technology. Background technique [0002] Lossy coding standards are widely used in the world to encode, compress and transmit images and videos. Specifically, image coding standards such as JPEG and video coding standards such as HEVC all adopt a block-based coding and compression framework: each frame of image is divided into independent coding blocks, transformed into the transform domain, and reduced by quantizing the transform domain coefficients. Transform domain coefficients, thereby reducing the bitrate. On the one hand, this quantization coding method determines the size of the code stream, and on the other hand, it also directly affects the reconstruction quality of the image. Using a large quantization step size will bring extremely low bit...

AI Technical Summary

Problems solved by technology

Nevertheless, since the texture reconstructed using an adversarial generative network is not consistent at the pixel level, the objective quality of the resulting decoded enhanced image is very low, especially in highly structured regions such as edges.

[0005] In summary, the existing technologies do not fully consider the visual characteristics of the human eye on different areas of the image (ie, edges and textures), and cannot achieve a balance between subjective quality and objective quality.

Images