Abstract:

Based on the research of image structural properties in DCT domain, DCT Domain-based Structural Similarity (DCTSSIM) was proposed. The method takes DCT domain information as the structural information and weights the DCT coefficients by different sensitivities of human visual system. Experimental results illustrate that, compared with SSIM and Peak Signal to Noise Ratio (PSNR), the proposed method is more consistent with the subjective assessment of human beings for JPEG2000 and JPEG images, thus can be used to describe the visual perception of image effectively.

Key words: information processing, image quality assessment, structural similarity (SSIM), human visual systems (HVS), discrete cosine transform based structural similarity(DCTSSIM)