EXPLOITING GLOBAL AND LOCAL INFORMATION FOR IMAGE QUALITY ASSESSMENT WITH CONTRAST CHANGE

Image quality assessment (IQA) has undergone a booming period during the last decade. Contrast change, being an important type of visual alteration for images, has not been seriously treated yet in the current IQA research. To address this problem, in this paper we propose a new reduced-reference (RR) IQA metric for contrast-changed images by exploiting local similarity information and global statistics information (LAGSI). In order to verify the effectiveness of the proposed algorithm, we test LAGSI with a large number of competitors on the recently introduced CID2013 database as well as contrast change related subsets from TID2008 and CSIQ databases. Experimental results demonstrate the superiority of our approach over classical and state-of-the-art IQA metrics.