Deep Learning for Image-based Cervical Cancer Detection and Diagnosis — A Survey

Abstract

Cervical cancer is the fourth most common type of cancer found in females with a record of 570,000 incidences and 311,000 deaths in the year 2018 worldwide. It is caused by a virus known as Human Papilloma Virus (HPV). Screening if done early can reduce this prevalence. However, manual screening methods are not efficient in the detection of cervical cancer as a result of some factors. This, however, results in misdiagnosis and over-treatment. Therefore, researchers proposed screening cervical automatically by using traditional and deep learning techniques. This paper aims to review past work that has been done particularly in the deep learning domain and discusses future directions in the automated detection of cervical cancer. It is believed that this will ensure proper diagnosis and could potentially reduce the prevalence of cervical cancer.