Image Quality Assessment Based on Wavelet Domain Structure Similarity

Aiming at some deficiencies of structural similarity(SSIM) model in assessing badly blurred or low bit-rate compressed images, a new image quality assessment algorithm based on wavelet domain structure similarity (WDSSIM) is proposed. It decomposes the whole image into subband images with different scales and frequencies, and then compares the structural similarity of each subband image between the original and degraded image, and weights the results according to the contrast sensitive function(CSF) properties of human visual system. Experimental results demonstratd that the WDSSIM algorithm could keep better cousistency with subjective perception for wider kinds of images.