کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
534952 | 870307 | 2008 | 10 صفحه PDF | دانلود رایگان |

In this paper, a novel framework combining the Chan–Vese active contour segmentation model with topology control mechanism is proposed. The novel framework can control active contours to capture specified objects in the applied images. Geometric active contours implemented by the traditional level set method can adaptively split and merge, but it is difficult to handle the topological constraints of the segmentation results. The proposed framework provides the users with previously set topologic constraints to control active contours’ splitting behavior for segmentation purposes. The Euler number and numbers of the connected regions of image objects are the topologic constraints implemented in the topology control mechanism. Some segmentation results of synthetic images demonstrate the effectiveness of the topological control algorithm. The segmentation results of the medical images show that the proposed framework provides convenience and flexibility to segment objects.
Journal: Pattern Recognition Letters - Volume 29, Issue 4, 1 March 2008, Pages 537–546