点击次数：

论文类型：期刊论文

发表时间：2015-08-10

发表刊物：Journal of Information and Computational Science

收录刊物：EI、Scopus

卷号：12

期号：12

页面范围：4557-4564

ISSN号：15487741

摘要：In this paper, we propose a novel color transfer method based on the saliency detection. By means of saliency maps obtained via saliency detection, the source and the target images are segmented into the foreground and the background regions, respectively. Then, we utilize the visual attention of images to match the regions between the two images. The foreground region belongs to the source image corresponds to the foreground part of the target one, therefore the source's background is naturally matched to the target's background. With the purpose of automatically transferring the color from regions of the source to the corresponding parts belong to the target in a perception aligned manner, we propose a saliency value based weighted color transfer model. It can efficiently prevent color mixing problems and successfully preserve the visual consistency of the results. Experimental results demonstrate that the proposed method performs favourably against most of the existing methods. Copyright ? 2015 Binary Information Press.