Morphologically restricted B-snake model for clustering it cell image segmentation

Abstract

In order to separate the clustering cells that frequently appear in cell images, an automatic segmentation algorithm is presented. The algorithm generates a modified "mlhear distaRce image," ad then ases active contour model te fincl cell hmtkrka. An iRIpreved iterative msion method, which uses a dynamic structure rather than a fixed one to produce the distance image, is proposed in the algorithm. A novel B-spline active contour model, whose energy depends not only on the image itself but also on the nonlinear distance image, is then presented in this paper. Initiated via morphological operation, the model would finally detect the cell boundaries. Experimental results show that our algorithm is effective for automatic segmentation on clustering cell images.