The provided research paper introduces ECgMLP, a novel deep learning model leveraging a gated multi-layer perceptron architecture, specifically designed for the automated and enhanced diagnosis of endometrial cancer from histopathological images. The study details the model's development, incorporating image preprocessing techniques like normalization and denoising, along with a watershed algorithm for region segmentation and photometric augmentation to improve data diversity. Through rigorous ablation studies and performance evaluations, ECgMLP demonstrates superior accuracy in classifying endometrial tissue compared to existing methods and other deep learning models, suggesting a significant advancement in computer-aided endometrial cancer diagnosis. The research highlights the potential of this approach to improve clinical workflows and patient outcomes through early and precise detection.

Sources: https://www.sciencedirect.com/science/article/pii/S2666990025000059