A full-reference image quality assessment method based on masked texture features

Abstract

The full-reference image quality evaluation method based on masked texture features disclosed by the present invention belongs to the technical field of image processing and image quality evaluation. First, the color space conversion is performed on the reference and distorted images, and secondly, the gradient magnitude and gradient direction features of the reference image and the distorted image are extracted. And calculate the image gradient information similarity, then calculate the texture feature similarity and color difference, and count the mean and standard deviation respectively to form a 6-D feature vector, build a regression model according to the random forest to fuse the feature vector and subjective MOS value, and Carry out training; finally extract the 6‑D feature vector of the image to be tested and input it into the trained regression model to complete the objective image quality evaluation. The evaluation method disclosed in the present invention adopts three different similarity features, uses random forest to establish a regression model, realizes full-reference image quality for high-precision objective evaluation, and can maintain high consistency with human visual characteristics.

Description

technical field [0001] The invention belongs to the technical field of image processing and image quality evaluation, and relates to a full-reference image quality evaluation method based on masked texture features. Background technique [0002] With the advent of the era of big data, more and more images are shared on the Internet. As an important carrier for people to obtain information and communicate, digital images are gradually changing people's lifestyles. With the large increase in data size, it also brings great challenges. During the process of image collection, storage, transmission and processing, a certain degree of distortion may occur. Therefore, how to effectively process and transmit images and accurately evaluate image quality has become an urgent research problem. [0003] In recent years, full-reference image quality assessment algorithms and corresponding devices are widely used in various image processing systems to optimize parameters, so full-refere...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a full-reference image quality evaluation method based on masked texture features, which solves the problem that the existing evaluation methods cannot accurately reflect the masking effect of human vision and ignore the influence of complex factors such as physiology and psychology on human vision. Problem, the present invention establishes a model by calculating the feature similarity of the reference image and the distorted image to achieve accurate quality evaluation of the distorted image

Images