bioimagesUnknownZixuan Panhttps://www.ebi.ac.uk/biostudies/studies/S-BIAD2133bioimagesSpecimenStudy ComponentBiosampleorganisationAssociationsAnnotationImage acquisitionYiyu ShiJianxu ChenMatthias GunzerJustin SonneckDennis NagelZixuan PanAnja HasenbergfalseAutoQC-Bench: A Diffusion Model and Benchmark for Automatic Quality Control on High-throughput MicroscopyThis dataset supports AutoQC, an automated quality control toolbox for high-throughput microscopy. It contains over 8,000 brightfield cell migration images, primarily high-quality training data, plus a curated subset with annotated examples of common image anomalies and reference masks. AutoQC uses a reconstruction-driven diffusion model trained only on normal images to detect diverse artifacts without labels. The dataset and models enable robust, scalable QC workflows for biomedical imaging. In addition to the raw data and annotations, we provide "normal_train.csv" and "normal_val.csv", listing the normal frames for training and for validation during training. We also provide "test_test_foldX.csv" (X = 0–4) and "test_val_foldX.csv", which define the fold-specific test and calibration spl2025-07-05T00:00:00Z2026-04-30T11:51:33.534Z2025-07-03T20:48:03.416ZS-BIAD2133