Image Quality Assessment in Reversible Data Hiding with Contrast Enhancement

Abstract

In this paper, image quality assessment (IQA) in reversible data hiding with contrast enhancement (RDH-CE) is studied. Firstly, the schemes of RDH-CE are reviewed, with which image contrast can be enhanced without any information loss. Secondly, the limitations of using the peak signal-to-noise ratio (PSNR) to indicate image quality in the scenario of RDH-CE are discussed. Subsequently, three no-reference IQA metrics and four metrics specially designed for contrast-changed images are adopted, in addition to PSNR and structural similarity (SSIM) index. By using these metrics, the evaluation results on the contrast-enhanced images generated with two RDH-CE schemes are obtained and compared. The experimental results have shown that the no-reference IQA metrics, the blind/referenceless image spatial quality evaluator (BRISQUE) for instance, are more suitable than PSNR and SSIM index for the images that have been enhanced by the RDH-CE schemes. Furthermore, how to use the suitable IQA metrics has been discussed for performance evaluation of RDH-CE schemes.