Weakly Supervised Learning using Attention gates for colon cancer histopathological image segmentation

Amina Ben Hamida, Maxime Devanne, Jonathan Weber, Caroline Truntzer, Valentin Derangère, François Ghiringhelli, Germain Forestier, Cédric Wemmert

Abstract: Highlights•A novel training strategy is introduced for sparsely annotated histopathological data.•Both UNet and Att-UNet are used for colon cancer Whole Slide image segmentation.•Enhanced lighter Att-UNet models are introduced for Whole Slide Image segmentation.