AN APPROACH FOR OVERLAPPING CELL SEGMENTATION IN MULTI-LAYER CERVICAL CELL VOLUMES

We propose a new method for detecting and segmenting overlapping cells in Pap smear images of multi-layer cervical cell volumes. This is a critical step in automating the analysis of cervical cells. This method first finds the nuclei inside the image using an iterative thresholding approach, then segment the location of overlapping cells, called cell clumps. The provided images from different focal are then used to segment the cytoplasm corresponding to each nucleus. The algorithm is evaluated using the dataset provided in the second overlapping cervical cytology image segmentation challenge at ISBI 2015. The results are presented based on two-fold cross validation on the training data which is provided with ground truth.