A Methodology for Deriving Objective Image Comparison Metrics

M.V. Le, P.D. Le, and B. Srinivasan (Australia)


image comparison, image quality measure, image similarity metric, image noise evaluation, wavelet representation.


In this paper, we propose a methodology for deriving objec tive image comparison metrics. It is based on the wavelet representation of images. In principle, a set of wavelets based noise measures is employed as the basis for deriving objec tive image comparison metrics. To test the validity of our ap proach, we have developed two image comparison metrics: an image quality measure and an image similarity metric. The experimental results indicate that our proposed method ology is valid. The effectiveness of our approach is that we can deduce more robust and reliable objective image compar ison metrics by integrating more reliable image noise mea sures based on wavelet representation.

Important Links:

Go Back