Abstract: Image quality index is a critical factor to evaluate the reconstruction quality of super-resolution methods.However,the existing image quality assessment methods usually require a higher resolution image as the reference image,which can not be satisfied in many actual super-resolution reconstruction applications.Taking into account the facts that human eyes are more sensitive to the relative luminance,and the sharpness of image edges is proportional to the subjective image quality,a no-reference SVD-based image quality assessment method utilizing the corresponding mean-removing images,gradient images,and gray images has been proposed in this paper.In this method,the image to be evaluated is divided into several overlapped patches first.Then,based on SVD,sub-image quality indexes between the low resolution images and the reconstructed super-resolution image for each patches are evaluated respectively.Finally,all sub-image indexes are integrated into one total image quality index,called SSQI index.Several experimental results using both simulated and real images demonstrate that the SSQI index is an efficient and stable index for evaluating the quality of the reconstructed super-resolution image without any higher resolution reference.It can also give an objective evaluation result consistent with human subjective assessment results.