download0 view122
twitter facebook

공공누리This item is licensed Korea Open Government License

Title
Segmentation of Regions of Interest Using Active Contours with SPF Function
Author(s)
Farhan Akram최광남김정헌이찬건
Publication Year
2015-01-31
Abstract
Segmentation of regions of interest is a well-known problem in image segmentation. This paper presents a region-based image segmentation technique using active contours with signed pressure force (SPF) function. The proposed algorithm contemporaneously traces high intensity or dense regions in an image by evolving the contour inwards. In medical image modalities these high intensity or dense regions refer to tumor, masses, or dense tissues. The proposed method partitions an image into an arbitrary number of subregions and tracks down salient regions step by step. It is implemented by enforcing a new region-based SPF function in a traditional edge-based level set model. It partitions an image into subregions and then discards outer subregion and partitions inner region into two more subregions; this continues iteratively until a stopping condition is fulfilled. A Gaussian kernel is used to regularize the level set function, which not only regularizes it but also removes the need of computationally expensive reinitialization. The proposed segmentation algorithm has been applied to different images in order to demonstrate the accuracy, effectiveness, and robustness of the algorithm
Keyword
segmentation; active contours; signed pressure force; medical image
Journal Title
Computational and Mathematical Methods in Medicine
Citation Volume
2015
ISSN
1748-670X
Files in This Item:
There are no files associated with this item.
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
URI
https://repository.kisti.re.kr/handle/10580/14438
Export
RIS (EndNote)
XLS (Excel)
XML

Browse