关键词: 无参图像质量评价, 图像功率谱, 小波变换

Abstract: Aiming at the feature of blurred images high frequency components, this paper proposes a no-reference image quality assessment method for Gaussian blurred images based on power spectrum. This method uses wavelet transform to extract image high frequency components, and constructs the evaluation index on the basis of the high frequency component spectrum. Experimental results show that the proposed evaluation index is better than PSNR and SSIM, whose objective quality scores have better consistency with the subjective quality scores.

Key words: no-reference image quality assessment, image power spectrum, wavelet transform