科研之友
微信
新浪微博
Facebook
分享链接
摘要
A novel Watershed segmentation approach is proposed in this paper. Firstly a review of medical image segmentation was presented where the advantages and drawbacks of the traditional watershed method are discussed. To resolve the existing problem, the proposed algorithm employs a threshold operation according to the range determined by the region of interest (ROI) and the modality of the medical image. And a "reslice" operation was employed to shrink the intensity of the image into a smaller range. Finally a resample operation is introduced to reduce the resolution of the image and this make watershed transform run in shorter time and less memory. The results indicated that the proposed method is efficient in solving the over segmentation problem and faster than classic watershed.
- 出版日期2006
- 单位上海交通大学
