CLEAR-IT, a framework for Contrastive Learning to Capture the Immune Composition of Tumor Microenvironments (software, GitHub mirror)

CLEAR-IT is a self-supervised (contrastive learning) framework for extracting cell-level features from multiplex tissue images to enable label-efficient cell phenotyping in tumor microenvironments. This software record archives the CLEAR-IT codebase used for computational analyses in the CLEAR-IT manuscript. The archived snapshot corresponds to Git tag v1.0.1 and is preserved for reproducibility. CLEAR-IT provides training and evaluation pipelines for self-supervised/contrastive learning and downstream cell phenotyping in multiplex imaging data. The repository includes source code, configuration files, experiment recipes, plotting notebooks, and citation/license metadata.Related research outputs are published separately as datasets: TNBC1-MxIF8 image dataset (DOI: 10.4121/126d8103-6de5-4493-a48e-5d529fef471e) and CLEAR-IT supplementary data (DOI: 10.4121/ebc792ad-4767-4aef-b8ff-ae653e901e3f).