Image processing for cervical pathology diagnosis using cervix’s polarized images

Abstract

Cervical pathologies are frequently occuring diseases and may affect women’s quality of life in many ways. These pathologies are curable with early detection and with a following suitable treatment plans. Colposcopy is a standard examination among screening methods which are used to early detect the abnormal lesions on cervix’s surface. Recently, studies about processing polarized image show ability to support diagnosis of the cervix. In this research, we use cervix’s polarized images and image processing algorithms to segment the blood distribution of Nabothian cyst and Trichomonas vaginalis infection. These results have the potential to provide underlying information of the cervix to support the diagnosis.