Segmentation of overlapping cytoplasm and overlapped areas in Pap smear images

• Published in: 2017 8th International Conference on Information, Intelligence, Systems & Applications (IISA) pp. 1 - 5
• Main Authors: Riana, Dwiza, Hidayanto, Achmad Nizar, Widyantoro, Dwi H., Mengko, Tati Latifah R., Kalsoem, Oemie
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2017
• Subjects: Cervical cancer; Feature extraction; Image edge detection; Image segmentation; Inflammatory Cells; Microscopy; Object recognition; overlapped areas; overlapped cell segmentation; Pap smear images
• DOI: 10.1109/IISA.2017.8316386

The ample of overlapping cells that complicate the slides reading and interpretation by pathologists has become a problem encountered in the conventional retrieval result. Another delicate problem is determining and analyzing the nucleus in the Pap smear of microcospic cells with inflammatory. A method for detecting overlapping cells and reducing the inflammatory cells to facilitate cell identification is required to solve the problems of overlapping cells and inflammatory cells. The proposed method is based on the fact that in isolated cells, the cell nucleus is located in the center of the surrounding cytoplasm. The first phase of the proposed method is the segmentation of the overlapping cells which consists of overlapping cytoplasm segmentation and overlapped area segmentation. The second phase of the proposed method is the overlapping cell separation and the last phase is the inflammatory cells ellimination and nucleus detection. The method has been evaluated on single cell images of conventional Pap smears, and the results are very promising.