Survey of Deep Learning Architectures for Ovarian Cancer Detection Using EfficientNetB0 and FPN

Abstract

Ovarian cancer is one of the leading causes of cancer-related mortality among women due to its asymptomatic nature in early stages and the lack of effective screening methods. Medical imaging techniques such as ultrasound, MRI, and CT scans are commonly used for diagnosis, but manual interpretation is often time-consuming and prone to variability. Recent advancements in deep learning, particularly convolutional neural networks (CNNs), have significantly improved automated ovarian cancer detection systems. This survey reviews recent developments in semantic segmentation and classification methods for ovarian cancer detection using architectures such as EfficientNetB0, Feature Pyramid Networks (FPN), and causal dilated convolutional neural networks. EfficientNetB0 enables efficient feature extraction from complex medical images, while FPN improves multi-scale feature representation for accurate tumor identification. Semantic segmentation models including U-Net and its variants are widely applied to identify tumor boundaries, whereas classification models distinguish between benign and malignant tumors. Deep learning systems have achieved diagnostic performance comparable to expert radiologists. The survey highlights important research trends such as hybrid architectures, attention mechanisms, and contextual feature learning. However, challenges including limited medical datasets, model interpretability, and poor generalization remain major concerns. Future research should focus on explainable AI, multi-modal data integration, and scalable clinical deployment of intelligent ovarian cancer detection systems.