Overfitting in Histopathology Model Training: The Need for Customized Architectures
This addresses the problem of overfitting for researchers and practitioners in medical imaging, particularly in histopathology, but is incremental as it builds on existing methods with domain-specific adaptations.
This study tackled overfitting in deep learning models for histopathology image analysis by showing that fine-tuning large-scale natural image models leads to suboptimal performance, and demonstrated that simpler, domain-specific architectures can achieve comparable or better results while minimizing overfitting on a public Oesophageal Adenocarcinomas dataset.
This study investigates the critical problem of overfitting in deep learning models applied to histopathology image analysis. We show that simply adopting and fine-tuning large-scale models designed for natural image analysis often leads to suboptimal performance and significant overfitting when applied to histopathology tasks. Through extensive experiments with various model architectures, including ResNet variants and Vision Transformers (ViT), we show that increasing model capacity does not necessarily improve performance on histopathology datasets. Our findings emphasize the need for customized architectures specifically designed for histopathology image analysis, particularly when working with limited datasets. Using Oesophageal Adenocarcinomas public dataset, we demonstrate that simpler, domain-specific architectures can achieve comparable or better performance while minimizing overfitting.