Automatic Contrast Enhancement Technology with Saliency Preservation
Ke Gu,Guangtao Zhai,Xiaokang Yang,Wenjun Zhang,Chang Wen Chen
DOI: https://doi.org/10.1109/tcsvt.2014.2372392
IF: 5.859
2014-01-01
IEEE Transactions on Circuits and Systems for Video Technology
Abstract:In this paper, we investigate the problem of image contrast enhancement. Most existing relevant technologies often suffer from the drawback of excessive enhancement, thereby introducing noise/artifacts and changing visual attention regions. One frequently used solution is manual parameter tuning, which is, however, impractical for most applications since it is labor intensive and time consuming. In this research, we find that saliency preservation can help produce appropriately enhanced images, i.e., improved contrast without annoying artifacts. We therefore design an automatic contrast enhancement technology with a complete histogram modification framework and an automatic parameter selector. This framework combines the original image, its histogram equalized product, and its visually pleasing version created by a sigmoid transfer function that was developed in our recent work. Then, a visual quality judging criterion is developed based on the concept of saliency preservation, which assists the automatic parameters selection, and finally properly enhanced image can be generated accordingly. We test the proposed scheme on Kodak and Video Quality Experts Group databases, and compare with the classical histogram equalization technique and its variations as well as state-of-the-art contrast enhancement approaches. The experimental results demonstrate that our technique has superior saliency preservation ability and outstanding enhancement effect.