Image quality measure based on curvelet transform

This paper presents a new image quality assessment based on the curvelet decomposition. The curvelet transform is introduced for its ability to represent the curves of images at different scales and directions. Unlike existing methods, the proposed image quality assessment takes into account the geometrical content of the image at different scales and directions. The proposed measure has been tested on LIVE database and was compared to other image quality assessments. The obtained results expressed in terms of correlation with subjective evaluation on the database demonstrate the efficiency of the proposed image quality metric.