Affiliations: [a] School of Electronic and Information Engineering, South China University of Technology, Guangzhou, China
| [b] School of Computer and Information Engineering, Henan Normal University, Xinxiang, China
| [c] School of Mathematics, South China University of Technology, Guangzhou, China
Correspondence:
[*]
Corresponding author. Zhiheng Zhou, School of Electronic and Information Engineering, South China University of Technology, Guangzhou, China. E-mail: [email protected].
Abstract: Some active contour models proposed based on intensity inhomogeneity are sensitive to initialization and cannot achieve ideal segmentation results for real images. An adaptive active contour model based on local bias field estimation and saliency is proposed in this paper. First of all, this model proposes an adaptive multi-local search algorithm, which avoids the initialization sensitivity by adaptively setting of the initial contour; Secondly, the local bias field is estimated by fusing the saliency map and fuzzy c-means clustering; Finally, the new bias field and the corrected energy fitting constant are used to define the new energy functional. The desired target object is obtained by minimizing the energy functional. The experimental results show that the segmentation accuracy of the model proposed in this paper is higher than that of the models participating in the comparison. The proposed model can not only avoid the interference of initialization and redundant information, but also segment images with intensity inhomogeneity effectively.
Keywords: Active contour model, intensity inhomogeneity, bias field, saliency map
