Comparison of nucleus and inflammatory cell detection methods on Pap smear images

Abstract

Detection and identification of cells in a Pap smear test is very important for determining cell abnormalities. Detection of cells becomes an important stage in early detection of cervical cancer. The presence of inflammatory cells in Pap smear images often results in the identification of nuclei in the early detection of cervical cancer in Pap smear image becomes difficult to do. The success of inflaming the inflammatory cells and detecting the nucleus will facilitate the process of identifying the nucleus. This will greatly help the development of information acquisition system technology and the classification of single Pap smear cell image for early detection of cervical cancer. The paper aims to compare three methods of nuclear detection and inflammatory cell in a single Pap smear image. The results of the comparison of nucleus and inflammatory cell in the test of data consisting of 84 images with inflammatory cells, showed that the K-Means and Bayesian classification method has not been able to accurately detect the nuclei and inflammatory cell, rather than the nuclei and inflammatory cells detection base on the combination of gray level thresholding method.