PROJECT TITLE :

Segmenting overlapping cervical cell in pap smear images - 2016

ABSTRACT:

Correct segmentation of cervical cells in Pap smear images is a crucial task for automatic identification of pre-cancerous changes in the uterine cervix. One in all the most important segmentation challenges is that the overlapping of cytoplasm, that was less addressed by previous studies. During this paper, we tend to propose a learning-primarily based method to tackle the overlapping issue with strong form priors by segmenting individual cell in Pap smear pictures. Specifically, we 1st outline the matter as a discrete labeling task for multiple cells with a appropriate cost operate. We tend to then use the coarse labeling result to initialize our dynamic multiple-template deformation model for additional boundary refinement on every cell. Multiple-scale deep convolutional networks are adopted to learn the diverse cell appearance features. Additionally, we tend to incorporate high level form info to guide segmentation where the cells boundary is noisy or lost due to touching and overlapping cells. We evaluate the proposed algorithm on 2 completely different datasets, and our comparative experiments demonstrate the promising performance of the proposed method in terms of segmentation accuracy.