Classification of cervical cancer from MRI images using multiclass SVM classifier

Abstract

Cervical Cancer is the abnormal growth of tissues in the lower, narrow part of the uterus (womb) called the Cervix which connects the main body of the uterus, to the vagina or birth canal. Cervical cancer is one of the most common types of cancer that can be seen in women worldwide. Early detection and proper diagnosis can prevent the severity level and reduce the death rates .In this paper, we have proposed an automated diagnosis system of cervical cancer using texture features and Multiclass SVM (Support Vector Machine) Classifier in MRI images. Initially the MRI images are pre-processed to remove undesirable noises and other effects. After pre-processing, the image is segmented by Region growing method to obtain the region of interest. Texture features are extracted from the segmented region. Almost 22 features are extracted at the region of a segmented area and then passed on to Multiclass SVM Classifier to detect if the given image is cancerous or not. The results of the proposed techniques provide effective results for classifying cancerous and the non-cancerous image.