Research On Stereoscopic Image Quality Assessment Algorithms Based On Human Visual Perception

With the rapid development of the stereoscopic image technology,the stereoscopic image has a wide application prospect in the fields of education,games,medical treatment and so on.But in the process of collection,transmission,storage and processing,the stereoscopic image quality will be reduced,which greatly limits its development.Stereoscopic image quality assessment(SIQA)can measure the stereoscopic image quality,and provide feedback and guidance for the related image processing algorithms to improve the image quality.SIQA is an indispensable part of stereoscopic image system;occupying an important position in the field of stereo image.Therefore,SIQA is a research hotspot in recent years.In this paper,a new SIQA algorithm based on cyclopean vision image features and visual attention model is proposed by analyzing the previous algorithms,and studying the characteristics of human visual system:Firstly,this paper introduces the SIQA system which includes subjective image quality evaluation,full reference,semi reference,no reference image objective quality assessment algorithm,the six indexes that validate the performance of objective image quality assessment.And this paper introduces the challenges which SIQA is facing.Secondly,because the purpose of objective image quality assessment is to design a SIQA algorithm consistency with human subjective feelings.to understand and simulate human visual system,this paper introduces the physiology of the human visual system,and visual mechanism,such as brightness,visual characteristics of the region of interest,contrast sensitivity,visual mask effect.Finally,based on the work of the previous chapters,a SIQA algorithm is proposed accounting for cyclopean image features and visual attention model.Firstly,the cyclopean images are generated by the left and right views accounting for binocular rivalry and contrast sensitivity.Meanwhile,the visual attention region of the cyclopean is extracted by a dual strategy visual attention model.The cyclopean images are divided into two categories,high quality images and low quality images according to the quality of cyclopean images.For the high quality cyclopean images,the saliency map is got by saliency detection,as for the low quality cyclopean images,the distortion region is detected by binocular just noticeable difference(BJND)model.Secondly,the structural,gradient magnitude,and phase congruency are utilized to generate the similarity of the reference and distorted cyclopean images.Finally,a weight is given to visual attention region.The experimental results indicate that the proposed algorithm achieves higher consistency with subjective quality assessment compared with the most related SIQA methods,and it could be widely used in multimedia communication systems for quality evaluation and service feedback.