Edge Entropy Motion Model based Quality Assessment for Low and Natural Videos

Video quality estimation plays an important role in various applications of video processing such as compression, restoration, printing, enhancement and watermarking. Now a days the field of objective quality evaluation gets more interest of researchers with affluent algorithms which is being recommended for this purpose. The Quality of the video is accessed by two ways Subjective and Objective. In subjective quality assessment metrics the quality of video is being estimated by human observers. In this humans judge the quality of distorted video. And in objective quality estimation the quality is assessed by quality metrics or algorithms. In this paper we evaluate the quality of the video by using objective quality assessment metrics. There are large numbers of objective quality assessment metrics like PSNR (Peak signal to noise ratio), MSE (Mean square error), RRED (Reduced reference entropic difference), Correlation and BLIINDS model metrics. We apply some distortion effects on video and calculate the quality on the basis of these metrics. We purpose a quality metrics (EMM) Edge entropy motion model. According to this metrics quality of video is assessed on the basis of edges of objects in the frame and the edges are extracted on the basis of gradients and gradients of image extracts the edges on color basis RGB. We take PSNR as a standard metrics and compare the results of BLIINDS model metric and EMM metric with PSNR. BLIINDS metrics gives results on the basis of the shape parameters values of each frame difference. In this we are trying to predict our purposed EMM metrics quality score is more accurate then BLIINDS metrics and its fluctuations is more close to PSNR as compare to BLIINDS.