Reduced-Reference Image Quality Assessment for Single-Image Super-Resolution Based on Wavelet Domain

Image quality metric is a critical factor to evaluate the quality assessment of single-image super-resolution (SISR) methods. The existing image quality assessment methods can be classified into full-reference and no-reference methods. However, full-reference methods require ground-truth images which are not available in practice, no-reference methods heavily rely on datasets. In this paper, we propose a reduced-reference image quality assessment method for SISR. A single low-resolution image is used as the reference image. First, small patches are taken from images and then extract features by the wavelet transform. Next, the features are fitted into the generalized Gaussian model. Finally, the distance between the fitting parameters of the LR and SISR images is used as the quality measure of SISR. Compared with the no-reference methods, training is not needed in the proposed method which has low dependence on the size of datasets, and more efficient and robust.