No-reference image quality assessment method based on multi-scale analysis

Abstract

The invention discloses a multi-scale analysis-based non-reference image quality assessment method. The method can effectively assess the quality of depth image drawing technology synthetized images. According to the multi-scale analysis-based non-reference image quality assessment method of the invention, a condition that the reference image of a depth image drawing technology synthetized image is not available is considered; on the basis of priori knowledge that the depth image drawing technology can cause geometric distortion to damage the self-similarity property of a natural image, and damage degree tends to drop with the decrease of the size of the image, a non-reference image multi-scale analysis quality assessment model is built; and a mainstream database is adopted to test the performance of the model. The performance of the method has great advantages over the performance of an existing evaluation method. It is noteworthy that research on non-reference evaluation methods for depth image drawing technology synthetized images are few, thus, the method of the invention can fill the technical blank in this field and directs a new direction for the improvement of non-reference image quality evaluation algorithms in the future.

Description

technical field [0001] The invention belongs to an image quality evaluation method, and uses a multi-scale analysis method to establish a novel non-reference image quality evaluation method, which can effectively evaluate the image quality synthesized based on depth image rendering technology. Background technique [0002] The image synthesized based on deep image rendering technology can provide users with a free viewing experience, making it widely used in distance education, monitoring, medical treatment and entertainment and other fields. The principle of image synthesis based on depth image rendering technology is to synthesize a virtual image through adjacent images. During this process, geometric distortion is inevitably introduced. like figure 1 , Geometric distortion presents different visual effects compared with blurring and noise distortion. However, there is currently a lack of research on quality assessment methods for images synthesized based on depth image...

AI Technical Summary

Problems solved by technology

However, there is currently a lack of research on quality assessment methods for images synthesized based on depth image rendering techniques.

Most of the existing image quality assessment algorithms are based on reference images. Because the images synthesized based on depth image rendering technology are virtual images, their reference images cannot usually be obtained, so the existing methods are difficult to apply

Images