An Objective Evaluation Method of Asymmetric Multi-distortion Stereo Image Quality

Abstract

The present invention discloses an asymmetric multi-distortion stereo image quality objective evaluation method. The method is characterized in that: in the test phase, according to image feature dictionary tables of a local phase image and a local amplitude image under different distortion types constructed in the training phase, a sparse coefficient matrix of each sub-block in the local phase image and the local amplitude image of the tested stereo image is obtained through optimization; through the sparse coefficient matrix, an image quality vector of each sub-block in the local phase imageand the local amplitude image under different distortion types is constructed in the training phase; and finally, by performing multi-distortion fusion, local global fusion, left-right viewpoint fusion and phase amplitude fusion on the sparse coefficient matrix and the image quality vector, an image quality objective evaluation value of the tested stereo image is predicted, and the image qualityobjective evaluation value maintains better consistency with the subjective evaluation value. According to the method disclosed by the present invention, without calculating image feature dictionary tables and image quality dictionary tables during the test phase, complex machine learning training processes are avoided.

Description

technical field [0001] The invention relates to an image quality evaluation method, in particular to an objective evaluation method for asymmetric multi-distortion stereoscopic image quality. Background technique [0002] With the rapid development of image coding and display technologies, the research on image quality evaluation has become a very important link. The goal of the research on the objective evaluation method of image quality is to keep consistent with the subjective evaluation results as much as possible, so as to get rid of the time-consuming and boring subjective evaluation method of image quality, which can automatically evaluate the image quality by computer. According to the degree of reference and dependence on the original image, the objective image quality evaluation methods can be divided into three categories: full reference (Full Reference, FR) image quality evaluation methods, partial reference (Reduced Reference, RR) image quality evaluation method...

AI Technical Summary

Problems solved by technology

At present, the existing no-reference image quality evaluation method uses machine learning to predict the evaluation model, but its computational complexity is high, and the training model needs to predict the subjective evaluation value of each evaluation image, which is not suitable for practical applications. There are certain limitations

Especially for the objective evaluation of asymmetric multi-distortion stereo image quality, the existing objective evaluation methods for single-view multi-distortion image quality or single-distortion stereo image quality evaluation methods cannot be directly applied. Therefore, how to construct a multi-distortion stereo The dictionary of image features, how to construct a dictionary that can reflect the quality of multi-distortion stereoscopic images, how to distinguish between the dictionary that reflects the characteristics of multi-distortion stereoscopic images and the dictionary that reflects the quality of multi-distortion stereoscopic images, between different distortion types, and between left and right Establishing connections between viewpoints is a technical problem that needs to be solved in the research of objective evaluation of asymmetric multi-distortion stereoscopic image quality

Images