A New Image Quality Assessment Metric Based on Contourlet and SVD

Abstract

The purpose of research on image quality assessment (IQA) is to find proper methods to measure the quality of images. The subjective evaluation score by HVS is usually used as a standard to be compared with. So, the more similar the measuring process is to HVS, the better the result should be. Motivated by this idea, contourlet transform and singular value decomposition are used in this article to establish an IQA metric because contourlet transform has the characteristics similar to HVS. Our IQA metric according to full-references is called CT-SVD. The new metric is tested on the image database TID2013 and the performance is compared with those of present metrics. It is shown that our CT-SVD metric reaches more consistency with the subjective image quality assessment.