Binary Classification of Cervical Cancer: A Comparative CNN Analysis

Cervical cancer poses a major global health challenge which emphasizes the importance of early detection for improved patient outcomes. This study conducts a thorough comparative analysis of six CNN architectures using the Herlev dataset. MobileNetV1 emerged as the standout performer with 90% accuracy, reinforced by robust precision (88%) and recall (83%). EfficientNetB0 showcases commendable precision (91%) and recall (90%), while ResNet-50 and DenseNet201 exhibited comparable accuracies of 84%. PeleeNet demonstrated lower performance with 74% accuracy and a 60% F1 score. The findings highlight MobileNetV1’s potential for accurate and efficient classification, contributing significantly to the ongoing endeavor of leveraging deep learning techniques for early cervical cancer detection, ultimately enhancing healthcare delivery and patient outcomes.