Cervical cancer diagnosis by deep learning using new pap smear dataset

Cervical cancer is one of the most common types of cancer in women all over the world. It is a type of cancer that can be controlled much better with early diagnosis. It should be checked with regular screening tests. The oldest and most common screening method for cervical cancer screening is the pap smear test. The analysis and interpretation of the Pap smear test is a time consuming, arduous and difficult task. Many automated systems have been developed to automate this process. The aim of this study is to provide a natural set of images collected from new, current and clinical standard data for use in studies of classification of cervical cells. For this purpose, a six-class data set consisting of 4939 data in accordance with the Bethesda system was created. ResNet50, VGG16, DenseNet121 and a CNN model were trained on this data set and the results were compared. The VGG16 model achieved the best result with an accuracy rate of 88%.