Real-time cell sorting with scalable <i>in situ</i> FPGA-accelerated deep learning
In situ FPGA deployment of knowledge-distilled CNNs achieves 14.5 µs inference for real-time label-free cell sorting with 40× faster detection-to-trigger time than SOTA, enabling clinical-grade lymphocyte classification via 5000-fold compression.
Authors
- Khayrul Islam (ORCID: https://orcid.org/0000-0003-2199-553X)
- Jianzhong Han (ORCID: https://orcid.org/0000-0003-3412-2068)
- Jian Huang (ORCID: https://orcid.org/0000-0002-1649-0038)
- Joshua Agar (ORCID: https://orcid.org/0000-0001-5411-4693)
- Seda Ogrenci
- Ryan F. Forelli (ORCID: https://orcid.org/0009-0001-9282-1150)
- Nhan Viet Tran (ORCID: https://orcid.org/0000-0002-8440-6854)
- Yaling Liu (ORCID: https://orcid.org/0000-0002-0385-6109)
- Deven Bhadane
Institutions
- Northwestern University (US)
- Fermi National Accelerator Laboratory (US)
- Lehigh University (US)
- Sichuan University (CN)
- Coriell Institute For Medical Research (US)
- Cooper Medical School of Rowan University (US)
- West China Hospital of Sichuan University (CN)
- Drexel University (US)
- Temple University (US)
Publication Details
- Journal
- Digital Discovery
- Published
- 2025-11-24
- DOI
- https://doi.org/10.1039/d5dd00345h
- Primary Topic
- Cell Image Analysis Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Science Foundation
- Northwestern University
- International Institute for Nanotechnology, Northwestern University
- Lawrence Livermore National Laboratory