Abstract:
Since contrast change is easy to make some distortions, such as illumination, color, etc., this paper proposes an image quality assessment method for measuring contrast change, called Contrast-change Image Quality Assessment (CCIQA). In the first stage, the proposed CCIQA method separates the images into two components, i.e., luminance and chrominance, Then, the corresponding luminance distortion factor is extracted by considering the intensity change and light-shade contrast change while the corresponding chrominance distortion factor is extracted based on the chrominance component similarity. Finally, an intensity weight map of the illumination component is exploited to guide the pooling process to obtain the final image quality score. The proposed CCIQA method is widely test on four commonly-used databases, TID2008, TID2013, CID2013 and CCID2014. Experimental results have shown that the proposed CCIQA method is able to achieve a high correlation with subjective prediction for contrast change, and outperforms multiple state-of-the-art image quality assessment methods.