This paper proposed a novel method for extracting skeleton of the objects with fuzzy boundary which are possibly aroused by inaccurate segmentation, based on Euclidean distance transform and the curve evolution theory, which integrates the advantages of topology accuracy and good noise elimination effect due to fuzzy boundary. In this paper, firstly, we introduce curve evolution theory for processing the origin objects with fuzzy boundary, then use the proposed method to extract skeleton. Since the distance transformation method will not guarantee the skeleton connectivity, the method firstly calculate the gradient of the distance transform, then obtaining a vector field. We mainly just detect the critical points inside the objects as for each critical point belongs to a local segmentation part of the objects, finally, we use searching the shortest path method to connect these critical points to obtain the whole skeleton, which can reduce data complexity and guarantee connectivity. The results demonstrate that the method is valid on accuracy and complexity. © 2013 Springer-Verlag.
CITATION STYLE
Zhang, Y., Li, L., & Yu, Y. (2013). A method for extracting skeleton of objects with fuzzy boundary. In Lecture Notes in Electrical Engineering (Vol. 212 LNEE, pp. 741–747). https://doi.org/10.1007/978-3-642-34531-9_78
Mendeley helps you to discover research relevant for your work.