Image Quality Assessment Method Based on Contourlet Transform and Generalized Gamma Distribution

This paper proposes a new image quality assessment approach based on Contourlet transform and generalized gamma distribution (GGD). The image is processed by Contourlet transform at first and the Contouelet coefficients are modeled by GGD. The parameter of GGD is estimated and the feature vector is formed by combining the parameter of GGD. The feature vector is fed to the support vector machine for training and testing. Experimental results on LIVE database show that our approach has good correlation with the subjective character of human vision.