Fast SSIM Index for Color Images Employing Reduced-Reference Evaluation

Abstract

Image quality assessment employing reduced-reference estimation approaches evaluate the perceived quality with only partially extracted features of the reference image. The primary aim of these approaches is to make objective evaluation flexible enough; accommodating the effect of any new distortion introduced in the image. Based on this concept, the paper proposes a fast approach for quality assessment of color images by modifying the SSIM index. The methodology involves sub-band decomposition of color images in wavelet domain for extracting statistical features. The computational complexity during estimation of features is reduced in this work by using the gradient magnitude approach. Noteworthy reduction in computational time is observed with the proposed index and the evaluation is also found coherent with full-reference and reduced-reference SSIM. © Springer International Publishing Switzerland 2014.