Automated segmentation of gallstones in ultrasound images

摘要：

Gallstones are solid particles formed from bile in the gall bladder. In this paper, we propose a technique to automatically detect Gallstones in ultrasound images, christened as, Automated Gallstone Segmentation (AGS) Technique. Speckle Noise in the ultrasound image is first suppressed using Anisotropic Diffusion Technique. The edges are then enhanced using Unsharp Filtering. NCUT Segmentation Technique is then put to use to segment the image. Afterwards, edges are detected using Sobel Edge Detection. Further, Edge Thickening Process is used to smoothen the edges and probability maps are generated using Floodfill Technique. Then, the image is scribbled using Automatic Scribbling Technique. Finally, we get the segmented gallstone within the gallbladder using the Closed Form Matting Technique.

关键词： Anisotropic Diffusion Unsharp Filtering NCUT Floodfill Technique Automated Scribbling Closed Form Matting

作者: Shivi Agnihotri Harsh Loomba Abhinav Gupta Vineet Khandelwal

作者单位: Department of Electronics & Communication Jaypee Institute of Information Technology University Noida, India

会议类型: 国际会议

会议名称: 2009 2nd IEEE International Conference on Computer Science and Information Technology(第二届计算机科学与信息技术国际会议 ICCSIT2009)

会议地点: 北京

会议语种:英文

页码: 56-59

在线出版日期: 2009-08-08（万方平台首次上网日期，不代表论文的发表时间）