A REVIEW OF DEEP LEARNING APPLICATIONS IN CERVICAL CANCER CELL DIAGNOSIS AND SCREENING

Tóm tắt

Cervical cancer ranks as the second leading cause of cancer-related mortality among women globally, claiming over 700 lives daily, with projections of 400,000 annual deaths by 2030. Early detection of cervical cancer and precancerous stages can lead to a full cure, but current screening methods like Pap smears and colposcopy suffer from high error rates due to human interpretation. Deep learning has become widely adopted as an effective tool in healthcare, capable of addressing complexities beyond the traditional artificial intelligence. To overcome manual screening limitations, computer-aided diagnosis systems using deep learning and machine learning are gaining traction, particularly for cervical cancer screening in underdeveloped regions where mortality rates are highest. This article reviews advanced deep learning techniques for analyzing cervical cytology and colposcopic images, focusing on classification and segmentation methods. It evaluates current deep learning algorithms in cervical cancer screening, discussing their potential to enhance diagnostic accuracy and accessibility. Additionally, the paper highlights ongoing research, challenges, and future directions, emphasizing deep learning’s role in reducing the global cervical cancer burden through improved screening solutions.