Classification of cervical cancer tissues using a novel low cost methodology for effective screening in rural settings

Cervical cancer is the second highest killer of women in Sub-Sahara Africa. This is due to unavailability of adequate screening methods and inability to pay where existing methods are available. This work addresses the issue of identifying and classifying cervical cancerous tissues for images taken from a standard camera. A cost effective method is developed for the imaging of cervical cancer in developing countries by analysing images from a standard camera. The image is enhanced using noise reduction and a Canny edge detector algorithm to find the intensity of the edges. Preliminary results obtained from clinical settings and images obtained are analysed by identifying the edges of the tissues of the cervix and classifying them based on the frequency of the edges. The results show above 90% classification accuracy. This shows that image analysis algorithm has the potential to successfully diagnose cervix cancer.