Effective Feature Selection for Pre-Cancerous Cervix Lesions Using Artificial Neural Networks

Since most common form of cervical cancer starts with pre-cancerous changes, a flawless detection of these changes becomes an important issue to prevent and treat the cervix cancer. There are 2 ways to stop this disease from developing. One way is to find and treat pre-cancers before they become true cancers, and the other is to prevent the pre-cancers in the first place. The presented approach uses precancerous images which are taken from a digital colposcope, and a set of texture and color features is extracted which includes low and high grade SIL (Squamous Interepithelial Lesion ) .After extracting, features are fed to a classifier, which could be KNN,RBF,MLP and Neuro-Fuzzy network and after training effective features are selected using UTA algorithm for each classifier individually. Finally, results come in a comparison table, show that the landa fourteenth, theta-x and together with Neuro-fuzzy classifier have the best overall performance. This approach has an acceptable and simple early diagnosis of cervix cancer and may have found clinical application.