Review of the Image Assessment Evaluation Algorithm Based on Local Information and Structure

Abstract

Image quality assessment (IQA) of an image is a vital research direction in the field of image processing. With the quickly developing evaluation methods, study of image quality evaluation method has a profound impact on image and video processing. At present, most of the image quality assessment methods are based on full reference. The traditional quality evaluation method based on structural similarity has the characteristics of narrow application range, unstable evaluation algorithm and so on. The difficulties of obtaining the original image and the difficulty of the reference model design directly incented the development of the semi reference IQA. In this context, a topic-based image quality evaluation process grounded on local information and structure information is proposed. Compared to the traditional image quality evaluation algorithm, the structure similarity has been improved by combining local information. The results show that the evaluation results of the method are more reasonable and stable, which provides a good reference for the post processing of the image.