Comparative Analysis of Transfer Learning Models for Breast Cancer Detection
Shahid Mohammad Ganie, Pijush Kanti Dutta Pramanik · 2024
Breast cancer is a significant health issue that leads to a high number of deaths in women worldwide. A successful course of therapy and better patient outcomes depend on early detection and diagnosis of breast cancer. Deep learning methodologies, specifically transfer learning, have recently been widely used in medical image analysis. This study investigates the comparative effectiveness of various transfer learning models for breast cancer detection. Five pre-trained models, namely VGG19, ResNet50, EfficientNetB4, MobileNetV2, and InceptionV3 are compared. These models are fine-tuned on a breast cancer image dataset comprising breast ultrasound scans. Their performances for 3-class classification are evaluated using seven metrics: accuracy, precision, recall, F1-score, support, AUC, and loss. This comparative analysis will help researchers and healthcare professionals choose the best transfer learning model for breast cancer detection by revealing their strengths and weaknesses.