anasanchezf/plate_effects

Multi-Source Domain Adaptation for Bioimaging Data (MSCDA-BioIm) MSCDA-BioIm is a biomedical microscopy benchmark for evaluating test-time and in-context domain adaptation under realistic batch effects. Built from the large-scale JUMP-CP dataset, it targets mechanism-of-action (MoA) classification using five-channel images of compounds associated with eight well-defined MoA classes. The dataset is organized by experimental batches and imaging… See the full description on the dataset page: https://huggingface.co/datasets/anasanchezf/plate_effects.