Research Of Video Quality Evaluation Method Based On Human Visual System

As computer network and multimedia technologies boom,digital video technology has been widely used in various fields.People have higher requirements on the quality of videos.Nevertheless,due to many constraints from real-world environments,videos may get distorted in a variety of ways during the transmission process.It is an urgent problem to filter high-quality videos and set video transmission.As an effective means to evaluate video quality,video quality assessment technology has been widespread concerned in recent years.Since the ultimate receivers of videos are humans,the assessment results of video assessment algorithms should be consistent with the subjective perception of human eyes.However,the traditional video quality assessment method does not combine the visual characteristics of the human eyes,so the assessment results are often inconsistent with the subjective perception of human eyes,which is difficult to provide correct guidance for video quality assessment.How to design a more accurate and efficient video quality assessment algorithm is the focus of current research.By combining the studies on the characteristics of human visual system,the following work has been conducted:1.An objective video assessment method considering inter-frame perceptual weighting is proposed by combining the persistence-of-vision effect of human eyes.The method first uses the JND model to judge the visibility of distortion,ignores invisible distortion,and implements perceptual weighting on the visible distortion in different areas within the frame by using the area division technology in combination with the attentional characteristics of human eyes,making the quality of a single frame more objective and accurate;performs inter-frame perceptual weighting based on the persistence-of-vision effect of human eyes,uses the inter-frame luminance difference to judge scene changing in the video sequence since the persistence-of-vision effect usually occurs when a scene is changed,the human persistence-of-vision effect will decrease the discrimination of human eyes during the action time,and when the video scene is changed,performs perceptual weighting on the affected frames;and finally,evaluates the performance of the algorithm by using the linear correlation coefficient and Spearman correlation coefficient,and compares it with the performance of traditional algorithm through experiments.The experimental results show that the algorithm can describe video quality objectively and perform well in materials with more dramatic scene changes.2.As an important component and attention-focusing area of a video,motion reflects the time domain characteristics of the video.In this paper,a video quality assessment method based on the classification of moving behaviors is proposed by combining the discrimination characteristics of human eyes.This method first conducts inter-frame perceptual weighting according to persistence-of-vision effect,uses motion detection technology to capture the moving objects in the scene,eliminates some small objects through the attentional characteristics of human eyes,then subdivides moving behaviors into the smooth type centered on motion perception and the flashing type centered on persistence effect according to the moving speed of the moving object in motion detection binary image results,and finally,further reflects the different processing ways and perception capabilities of human eyes for different moving types through the perceptual weighting on different moving behaviors.Database sample experiments show that the algorithm is of high accuracy.