SAR River Image Segmentation by Active Contour Model Inspired by Exponential Cross Entropy
- 50 Downloads
Utilizing the existing active contour models to achieve accurate segmentation of SAR river images is ineffective. To address this difficult, a novel active contour model inspired by exponential cross entropy is proposed. The external energy constraint term of the proposed model is defined inspired by exponential cross entropy. Then, the means of the pixel grayscale values inside and outside the curve are utilized to modify the external energy constraint term, which can improve segmentation performance. Moreover, the Dirac function is replaced by the edge magnitude function to accelerate the curve evolution, which can improve segmentation efficiency. The extensive experiments are performed on a large number of SAR river images and the results demonstrate that the proposed model outperforms the existing active contour models in terms of both segmentation performance and segmentation efficiency.
KeywordsImage segmentation SAR river image Active contour model Exponential cross entropy Edge magnitude function
This work is partially supported by the National Natural Science Fund of China under Grant 61573183, Key Laboratory of Yellow River Sediment of Ministry of Water Resources under Grant 2014006, Engineering Technology Research Center of Wuhan Intelligent Basin under Grant CKWV2013225/KY, State Key Laboratory of Urban Water Resources and Environment under Grant LYPK201304. Funding for Outstanding Doctoral Dissertation in NUAA under Grant BCXJ18-04, Postgraduate Research & Practice Innovation Program of Jiangsu Province under Grant KYCX18_0288.
- Caselles, V., Kimmel, R., & Sapiro, G. (1995). Geodesic active contours. In Proceedings of fifth international conference on computer vision (pp. 694–699).Google Scholar
- Li, C., Kao, C., Gore, J., & Ding, Z. (2007). Implicit active contours driven by local binary fitting energy. In Proceeding of IEEE conference on computer vision and pattern recognition (pp. 1–7).Google Scholar