AUTOMATIC DETECTION OF CERVICAL CANCER USING UKF AND ACM-AHP

Cervical cancer affects the area of the cervix and is the lower portion of women's abdomen and extends into the vagina, which does not contribute to any early stage diagnosis symptoms. Pap smear test is the manual cervical cancer screening process that leads to high false positives due to human error. The manual screening method is expensive and there are only few experienced pathologists available for the diagnosis of this test. Because of the unusual borders between nucleus and cytoplasm the segmentation between Pap smear slides is very difficult. In cellular structure, complexity increases as the cells collide with each other where the nucleus and cytoplasm borders are impossible to distinguish. The techniques aided by automated and semi-automatic computers that are used to section the nucleus and cytoplasm of cervical cells. In this paper, the Pap smear image has been enhanced by utilizing Upgraded Kaun Filter. The enhanced image has been segmented by Active Contour model. In this, the weight optimization issue has been resolved using Analytic Hierarchy Process. Hence, the effectual features are extracted from the segmented region which is useful for classifying the cancer by ENN-TLBO classifier. The results show that the method outperforms many of the existing algorithms in terms of the accuracy, sensitivity and specificity when applied to the Pap smear data set.