PARTITIONED GRAPH CUT SEGMENTATION ON HETEROGENEOUS CELL IMAGES
Authors: *Iruansi U, Oyebode K.O.
Heterogeneous cell images are currently posing great challenges in medical image segmentation. This is due to the fact that cells that constitute these images lack homogeneity. As a result, popular segmentation methods such as active contour, watershed, thresholding and graph cut do not perform well. In order to address this challenge, a partitioned segmentation of cell images is proposed using graph cut. First, cell images were broken into fairly homogeneous units and then graph cut segmentation was thereafter carried out on each individual partitioned unit. The segmentation results from these units were then merged. Experimental results show improved performance over traditional graph cut and state of the art models.
Affiliations: *Department of Computer Engineering, Faculty of Engineering, University of Benin, PMB 1154, Benin City, Nigeria
Keywords: Cell Segmentation, Graph Cut, Heterogeneous Cells, Partitioned Graph Cut, Segmentation
Published date: 2018/06/30