Research On No-Reference Quality Assessment Of Blur Image

With the rapid development of computer technology,electronic products have gradually occupied an important position in people's lives.Along with the emergence of images and videos,Before the image is acquired by the human eye,the image is compressed and is transmitted,which resulting in different degrees of distortion.Blurring is an important cause of image distortion.The quality assessment of blur image have gradually occupied an important position in the field of image processing.According to the degree of dependence on the original image,The quality assessment of blur image can be divided into three categories: full reference,reduced-Reference and no reference.Among them,the assessment result of the quality assessment of non-reference blur images has nothing to do with the original image feature,and it has stronger applicability and that structure is simpler.This paper studies the quality assessment of non-reference blur images.The main contents are as follows:(1)Since many existing assess algorithms do not consider image feature attributes such as structure information,correlation between pixels and pixels,and the like.When assessing two images with different contents,the algorithm will produce significant errors due to the difference in the extracted image information.In view of the above problems,this paper proposes a method of quality assessment of blur image based on fusion.The method uses the probability of blur detection to measure the degree of image blur,and introduces the variance describing the image structure information into the assessment algorithm,the image structure information is described by using the standard deviation of the two images obtained after the re-blurring.Finally,the assessment index is obtained by the fusion of the probability of blur detection and the structural information of the image.At the same time,when calculating the variation of the standard deviation,the algorithm eliminates the pixel points with large variance,and effectively suppresses the influence of noise introduced by re-blurring filtering.The experimental results show that the algorithm can solve the error problem caused by the large difference in image content.(2)The traditional assessment algorithm concentrates on the calculation of the blur of the image edge,ignoring the important significance of human visual attention.When assessing an image with different blur degree of background and foreground.the assessment result is lower than the result of subjective evaluation due to the influence of the blurred background.Aiming at the problems of the above traditional algorithms,this paper proposes an algorithm of the quality assessment of blur image based on problem of blur background.The algorithm first preprocesses the image,extracts the salient region of the image with adaptive salient threshold,and calculates the probability of blur detection of two images after re-blurring.and Finally,the image quality is described by the change of the two indicators.The algorithm introduces image saliency into the algorithm of assessment,and adaptively calculates the ambiguity of the image's salient region according to the image scene,eliminating the influence of the blur background of image,and making the assessment results more in line with human visual characteristics.The experimental results show that the algorithm achieves good experimental results in the LIVE data set,and has better assessment performance when compared with the existing three traditional algorithms.