Image quality assessment based on included angle cosine and discrete 2-D wavelet transform

In this paper, a novel image quality assessment based on the characteristics of wavelet coefficients of images and included angle cosine is proposed. Firstly, the normal image and the images assessed are decomposed into several levels by means of wavelet transform respectively. Secondly, the approximation and detail coefficients of normal image are as the referenced sequences and the approximation and detail coefficients of the images assessed are as the comparative sequences respectively. And the included angle cosine values are calculated between the referenced sequences and the comparative sequences respectively. Moreover, image quality assessment matrix of every image assessed can be constructed based on the included angle cosine values and image quality can be assessed. The algorithm makes full use of perfect integral comparison mechanism of included angle cosine and the well matching of discrete wavelet transform with multi-channel model of human visual system. Experimental results show that the proposed algorithm can not only evaluate the integral and detail quality of image fidelity accurately but also bears more consistency with the human visual system than the traditional method PSNR.