Automatic Approach for Cervical Cancer Detection Based on Deep Belief Network (DBN) Using Colposcopy Data

Cervical cancer is one of the diseases with the highest mortality rate. In the world, cervical cancer is ranked as the fourth most dangerous disease. Based on these problems, this paper can be an alternative to help medical authorities in detecting cervical cancer with the help of the Computer-Aided Diagnosis (CAD) System. CAD System used has two processes, such as preprocessing and classification. Preprocessing is useful to improve the image so that it is easier to do the process of identifying features. Preprocessing used is greyscale, histogram equalization, and median filter. Preprocessing results will be formed into a vector matrix using the reshaping process. The final step is the process of classifying data using the Deep Belief Network method. The best accuracy results obtained from the identification process of cervical cancer using the DBN method is 84%. Based on the results of accuracy, is expected to help reduce the number of deaths from cervical cancer with early detection.