Quality objective evaluation method with no reference images based on multi-scale generative adversarial network

Abstract

The invention discloses a quality objective evaluation method with no reference images based on multi-scale generative adversarial network. Similar quality images corresponding to distorted images canbe generated through the multi-scale generative adversarial network, and similar quality images in different scales can undergo regression to obtain the image quality scores through convolution nervenetwork. The multi-scale generative adversarial network is trained, and similar quality images are generated for the distorted images through the full-reference image quality evaluation method, and the similar quality images are considered as a real data set for determining network. Three groups of similar quality images in different scales are considered as the data set, and the subjective evaluation score is used as the label, and the image quality score regression network is trained. The distorted image generates a plurality of similar quality images in different scales through the generative network, and generates the image quality score through the image quality score regression network. The invention is advantageous in that the integral distortion degree and the local distortion details are combined, and the quality score of the distorted images can be further determined, and the quality of distorted images can be embodied in a more comprehensive and more accurate manner.

Description

technical field [0001] The invention relates to a non-reference image quality objective evaluation method based on a multi-scale generation confrontation network, which belongs to the technical field of digital image processing. Background technique [0002] Image quality evaluation plays an important role in algorithm analysis and comparison and system performance evaluation in image processing systems. In recent years, with extensive research in the field of digital images, researchers have paid more and more attention to the study of image quality evaluation, and proposed many indicators and methods for image quality evaluation. [0003] From the perspective of whether there is human participation, image quality evaluation methods can be divided into subjective evaluation methods and objective evaluation methods. Subjective evaluation uses people as observers to evaluate images subjectively, and strives to truly reflect human visual perception; objective evaluation metho...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problem of poor performance of the existing no-reference image quality evaluation, the present invention proposes a no-reference image quality evaluation method, which uses a generative adversarial network to generate multiple similar quality maps of different sizes for a distorted image, and these similar quality maps are passed through Convolutional Neural Networks for Regression to Get No-Reference Image Quality Scores

Images