Self-supervised Learning for Large-Scale Batch-Effect-Aware Quantitative Phase Imaging Cytometry
We introduce CytoMAD, a self-supervised deep learning strategy that synthesizes single-cell quantitative phase images and simultaneously distills cellular features from batch variations for consistent analysis across datasets. This approach enables diverse large-scale label-free imaging cytometry applications, including drug screening and tumor heterogeneity analysis.
Authors
- Kevin K. Tsia (ORCID: https://orcid.org/0000-0002-6394-9657)
- Dickson M. D. Siu (ORCID: https://orcid.org/0000-0002-1598-7253)
- Jcm Ho (ORCID: https://orcid.org/0000-0003-4499-5284)
- Michelle C. K. Lo (ORCID: https://orcid.org/0000-0002-4452-5525)
- Michael K. Y. Hsin
- Maximus C.F. Yeung (ORCID: https://orcid.org/0000-0001-9942-0109)
Institutions
- Chinese University of Hong Kong (HK)
- University of Hong Kong (HK)
Publication Details
- Published
- 2024-11-10
- DOI
- https://doi.org/10.1109/ipc60965.2024.10799912
- Primary Topic
- Digital Holography and Microscopy
- Type
- article
- Field-Weighted Citation Impact
- 0.00