Segmentation of Overlapping Areas on Pap Smear Images with Color Features Using K-Means and Otsu Methods
Dwiza Riana, Herdian Tohir, Achmad Nizar Hidayanto · 2018 Third International Conference on Informatics and Computing (ICIC) · 2018
Separation of overlapping areas on overlapping Pap smear image is still a tough thing to do. In addition to the process of identification of overlapping cells which is still difficult to do, sometimes there are components in the cells that complicate the process such as the presence of inflammatory cells that resemble the nucleus though the process of overlapping cell identification on Pap smear image is needed in early detection of Pap smear test to recognize a cell that is healthy or classified as cervical cancer cells. This study aims to segment the overlapping areas of Pap smear image by utilizing the color features of the cell. Stages of identification and separation of the area use K-means method combined with Otsu method. The results obtained in some cells of overlapping areas can be perfectly separated while in some other cells the results have not been encouraging and still require further research, especially to overcome the existence of inflammatory cells that complicate the separation process.