Abstract—a good segmentation result depends on a set of “correct” choice for the seeds. When the input images are noisy, the seeds may fall on atypical pixels that are not representative of the region statistics. This can lead to erroneous segmentation results. In this paper, an automatic seeded region growing algorithm is proposed for cellular image segmentation. First, the regions of interest (ROIs) extracted from the preprocessed image. Second, the initial seeds are automatically selected based on ROIs extracted from the image. Third, the most reprehensive seeds are selected using a machine learning algorithm. Finally, the cellular image is segmented into regions where each region corresponds to a seed. The aim of the proposed method is to automatically extract the Region of Interests (ROI) from the cellular images in terms of overcoming the explosion, under segmentation and over segmentation problems. Experimental results show that the proposed algorithm can improve the segmented image and the segmented results are less noisy as compared to some existing algorithms.

Index Terms—Image segmentation, Seeded region growing, Machine Learning, Leaking problem.

M. M. Abdelsamea is with Mathematics Department, Faculty of Science, Assiut University, Assiut, Egypt (e-mail: mm_abdelsamea@yahoo.com).



Cite: M. M. Abdelsamea, "An Enhancement Neighborhood Connected Segmentation for 2D-Cellular Image," International Journal of Bioscience, Biochemistry and Bioinformatics vol. 1, no. 4, pp. 256-260, 2011.