Segmentation and Classification of Cervical Cancer

Uterine cancer comes in second place after breast cancer affecting a large majority of women, especially in poor countries, given the lack of programming for the usual manual/visual screening to detect this cancer, it is in most cases proving insufficient because it is detected in advanced stages, thus the need to produce effective and reliable early detection methods. In this perspective, our work is interested in developing a tool to assist in the diagnosis and early detection of cervical cancer based on the interpretation of MRI images of the cervix, the proposed system takes place in three stages: 1-Pretreatment to remove the noise from the image in general; we opted for the K-means method; 2- Segmentation step in which we used the method of growing regions, 3- Classification or decision step that consists through inferences rules deduced from the FIGO classification to decide which stage it is. The results obtained are satisfactory and demonstrate the effectiveness of our approach to detecting the stage of cervical cancer.