点击次数：

论文类型：期刊论文

发表时间：2017-02-01

发表刊物：JOURNAL OF MEDICAL IMAGING AND HEALTH INFORMATICS

收录刊物：SCIE、Scopus

卷号：7

期号：1

页面范围：247-250

ISSN号：2156-7018

关键字：Visible Human; Matting; Closed-Form Solution; Distance Transform

摘要：In this paper, a new solution to hyper scale serialized visible human slices is proposed. A closed -form solution and distance transform strategy are synthetically utilized to solve the problem of low -quality image and non applicability in the ROI matting and seed point extraction procedure. Closed -form solution calculates a globally optimal result of foreground opacity. It bases on an assumption of a smooth relation between foreground and background colors. To meet the need of full automatic segmentation, a new seed point extraction method based on distance transform is proposed. The segmented image is transformed to be a distance matrix. Some stable points in this matrix are chosen as seed points. This solution is fine and smooth but with less user input. The sequential segmented ROI images are satisfactory and have many application prospects.