NO REFERENCE IMAGE QUALITY ASSESSMENT DEPENDING ON YCbCr AND L*u*v*

Abstract

Image quality assessment is a complicated and hard process, because of its independent on reference image, this process is done by some no reference metrics. In this paper the quality of images were assessed by a no reference metric, this metric is (The entropy of first derivative (EFD)), also Gaussian blurring were assumed as a distortion, and to estimate the efficiency of the no reference metric in assess the quality of colored images it is necessary to compare the results with the results of some reference metrics, we depended on (Normalize mean square error (NMSE) and structural similarity index measurement (SSIM)) as a reference metrics.Furthermore, two color spaces were applied (YCbCr and L*u*v*) to find the best color space to working on. In the first step of the work the selected images were distorted by Gaussian blurring then the quality of the distorted images were measured by the reference metrics, this is done in the two color spaced were used, after that the no reference metric were applied to assess the quality of the distorted images, finally, the results were compared with the results of the reference metrics by plotting and by finding the correlation coefficients between the carves of reference and no reference metrics to find the best color space to working on.