Tchebichef moment based image quality measure

Image quality measures try to quantify a visual quality or analogically, an amount of distortion in a given picture. This thesis work proposes a moment based image quality assessment metric that automatically assesses the quality of an image in the discrete orthogonal moments domain. In this work, the discrete orthogonal moment namely, a Tchebichef moment is chosen and quality measure is developed. The proposed quality metric is applied to the standard images in the IVC and A57 database, to assess its performance. The conventional quality measures such as MSE, PSNR, SSIM, MS-SSIM, UQI, VSNR and WSNR are also computed. Then the performance of the proposed quality metric is evaluated by comparing them with the conventional quality measures. The results portray that the proposed quality metric performs better than WSNR and VSNR metric in some cases. Also the proposed metric works well for all images with certain types of distortions.