Tallinn University of Technology

25 August 2025 at 1:00 PM
Fariha Afrin, "Microfluidic Droplet Detection, Classification and Quality Assessment for Embedded Flow Cytometry Systems"

Supervisor: Professor Yannick Le Moullec, Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia

Co-supervisor: Senior Research Fellow Tamás Pardy, Thomas Johann Seebeck Department of Electronics, Tallinn University of Technology, Tallinn, Estonia

Opponents: 

  • Assistant Professor Luca Reggiani, Dipartimento di Elettronica, Informazione e Bioingegneria (DEIB), Politecnico, Milano, Italy
  • Associate Professor András József Laki, Faculty of Information Technology, Pázmány Péter Catholic University, Budapest, Hungary 

Imaging flow cytometry (IFC) is a tool that analyze images of thousands of cells as they flow through tiny channels. This helps research in cancer, immunology, and cell biology. However, until now, this technology needed bulky optics equipment and computers with significant processing power, making it confined to well-equipped laboratories.  
To bring such technology to places outside traditional laboratories by using small, affordable optics and computing devices is challenging. This thesis addresses the challenges, resulting in three key advances to enable cell analysis work on small, portable devices such as Raspberry Pi: 

  1. Image Capture. A new lightweight method that captures clear images of fast-moving cells without needing expensive hardware. It rapidly and successfully counts microfluidic droplets not only on standard computers, but also on small computing devices. 
  2. Image Quality Assessment. A specialized system designed for resource-constrained devices that quickly determines if cell images are good enough for further analysis. This ensures reliable results with minimal computing resources, making quality control possible in the field. 
  3. Ultra-Efficient Analysis. An analysis system that rapidly classifies microfluidic droplets while maintaining high accuracy on small computing devices. The research also explored specialized hardware accelerators that further reduce processing time, though with some trade-offs in classification performance. 

Through careful system design and optimization, complex biological analysis can be performed on small, affordable computing devices. By optimizing for edge devices, it is possible to bring sophisticated cell analysis closer to the point of need, without having to send data to cloud servers or requiring expensive equipment. In turn, this opens possibilities for point-of-care diagnostics in clinics without lab facilities, field research in remote locations, and affordable cellular analysis in resource-limited environments. 

Join the public defence in MS Teams