¹Physical Engineering College, Zhengzhou University, Zhengzhou 450001, China
²Ningbo Key Lab of DSP, Zhejiang Wanli University, Ningbo 315100, China

---

Abstract—Active contour method is widely used in the image processing field. Recently, it has been used in object segmentation and has attracted great attention. However, most of the existing object segmentation methods based on active contours are complex and time-consuming. They cannot be used in some real-time applications. Hence, in this paper, a fast and efficient moving object segmentation algorithm based on an improved active contour model is proposed. Firstly, moving regions are detected based on frame differences. Then their boundaries are extracted and used as the coarse objects’ contours for the next segmentation based on active contour model. Thirdly, an improved active contour method is introduced to extract the final objects with the coarse objects’ contours as initial values. Experiments are implemented and the results show that the proposed algorithm can segment and track moving targets quickly and correctly.

Index Terms—Moving object segmentation, frame difference, active contour model, fast algorithm.

[PDF]

Cite: Hui-min Zhang, Zhi-yong Duan, Zhong-jie Zhu, Yuer Wang, "Fast Moving Object Segmentation Based On Active Contours," Journal of Computers vol. 7, no. 4, pp. 863-869, 2012.