Optical/Microscopy

2022 | Self-supervised driven consistency training for annotation efficient histopathology image analysis

 150 150 MaLMIC - Machine Learning in Medical Imaging Consortium

In this work, we propose two novel strategies for unsupervised representation learning in histology: (i) a self-supervised pretext task that harnesses the underlying multi-resolution contextual cues in histology whole-slide images to learn a powerful supervisory signal for unsupervised representation learning; (ii) a new teacher-student semi-supervised consistency paradigm that learns to effectively transfer the pretrained representations to downstream tasks based on prediction consistency with the task-specific unlabeled data. We carry out extensive validation experiments on three histopathology benchmark datasets across two classification and one regression based tasks.

2022 | Self supervised contrastive learning for digital histopathology.

 150 150 MaLMIC - Machine Learning in Medical Imaging Consortium

Short Description: We apply a contrastive self-supervised method to digital histopathology. A large-scale study with 57 histopathology datasets without labels was conducted. Our study focuses on differences between natural-scene and histopathology images. We find combining multiple multi-organ datasets improves task performances. Using more pretraining data has diminishing returns after around 50,000 images.
Modality: Optical/Microscopy
Organ: Prostate, Breast, Colorectal
Disease: Cancer

bottom of publications

 150 150 MaLMIC - Machine Learning in Medical Imaging Consortium