SEENS: Nuclei segmentation in Pap smear images with selective edge enhancement

Abstract Accurate nuclei segmentation, as an indispensable basis and core link for multi-cell cervical image analysis, plays important role in automatic pre-cancer detection. However, poor quality due to the uneven staining, complex backgrounds overlapped cell clusters poses a great challenge segmentation. In this paper, we propose new Selective-Edge-Enhancement-based Nuclei Segmentation method (SEENS). proposed method, selective search is integrated with mathematical operators segment whole slide images into small regions of interest (ROI) while automatically avoiding repeated segmentation well eliminating non-nuclei regions. addition, edge enhancement based on canny operator morphology presented extract information weight enhance nucleus selectively. As result, enhanced ROI then segmented by Chan–Vese model higher accuracy. We evaluate our 18 total 395 nuclei. Experimental results demonstrate that SEENS achieves accuracy Notably performs particularly better low-contrast scenarios than baselines.