CellMap 2024 Segmentation Challenge

The CellMap Segmentation Challenge provides a comprehensive dataset collection aimed at advancing machine learning segmentation in electron microscopy. This release includes 289 meticulously annotated training volumes derived from 22 diverse eFIB-SEM datasets, encompassing over 40 unique organelle classes. These datasets capture a wide variety of cellular structures across multiple biological contexts, offering a richly detailed resource for algorithm development. The annotations and dataset preparation protocols ensure high consistency and quality, setting the stage for significant progress in understanding cellular architecture and improving segmentation methods. All data are publicly available on OpenOrganelle, with a collection DOI for citation: https://doi.org/10.25378/janelia.c.7456966. Detailed instructions for accessing the data, participating in the challenge, and submitting predictions are available on the official GitHub repository. The white paper CellMap Segmentation Challenge describes the dataset collection, annotation protocols, and potential applications, providing a foundation for innovation in electron microscopy image analysis.