Protocol to annotate and automate single-cell instance segmentation on stimulated Raman histology using deep learning
Stimulated Raman histology (SRH) is a label-free optical imaging technique that can discern molecular components such as lipids and proteins at subcellular spatial resolution without histologic staining. Here, we present a protocol for labeling cells and training AI models for automated cell segmentation on SRH images acquired intra-operatively from neurosurgical cases. We describe steps to enable single-cell spatial analysis on SRH using ELUCIDATE, a web-based SRH cell annotation tool, and DetectSRH Python library.
Authors
- Akhil Kondepudi (ORCID: https://orcid.org/0000-0002-2643-7580)
- Edward Harake (ORCID: https://orcid.org/0000-0002-8865-1267)
- Lisa L. Walsh (ORCID: https://orcid.org/0000-0003-0311-5455)
- Xinhai Hou
- Abhishek Bhattacharya (ORCID: https://orcid.org/0000-0001-8663-9363)
- Cheng Jiang (ORCID: https://orcid.org/0000-0003-1759-4960)
- Todd Charles Hollon (ORCID: https://orcid.org/0000-0001-5987-6531)
- Asadur Chowdury (ORCID: https://orcid.org/0000-0002-5653-9721)
- Eric Landgraf
- Lin Wang (ORCID: https://orcid.org/0000-0002-6449-4930)
Institutions
- University of Michigan (US)
- NYU Langone Health (US)
- Michigan Medicine (US)
Publication Details
- Journal
- STAR Protocols
- Published
- 2025-11-28
- DOI
- https://doi.org/10.1016/j.xpro.2025.104221
- Primary Topic
- Spectroscopy Techniques in Biomedical and Chemical Research
- Type
- article
- Field-Weighted Citation Impact
- 0.00