Open AccessActive contours driven by novel fitting term for image segmentation
Author(s) -
Ni Kang,
Wu Yiquan
Publication year - 2017
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/el.2017.1531
Subject(s) - curve fitting , segmentation , term (time) , artificial intelligence , image segmentation , entropy (arrow of time) , reciprocal , computer science , computer vision , active contour model , cross entropy , pattern recognition (psychology) , algorithm , mathematics , physics , linguistics , philosophy , quantum mechanics , machine learning
Active contour model driven by novel fitting term is proposed for image segmentation in this letter. The novel fitting term contains two parts: one is L 1 fitting term, which describes the change of energy between inside and outside the curve; another is the reciprocal cross‐entropy fitting term, which can compute the local energy of every point on the curve. The reciprocal cross‐entropy fitting term is more robust than L 1 fitting term when the regional grey scale fluctuates greatly, and the detail direction of the curve can be controlled better by it. Experiments on several synthetic and real images have shown that the proposed model achieves better segmentation results and efficiency than the other models; in addition, it is insensitive to different initial contours.