Motion-aware imaging flow cytometry framework for robust counting and tracking of cells with non-uniform velocities in microfluidics

Imaging flow cytometry combines high-throughput cellular analysis with visualization but struggles with accurate quantification under non-uniform microfluidic flow conditions. To overcome this, we developed a motion-aware imaging flow cytometry framework integrating Kalman filter-based motion prediction and Hungarian algorithm-based cell association. The system captures sequential frames for cell tracking and counting, enabling detection of non-uniform velocity profiles and stationary adherent cells. Experimental validation using dual-view transport of intensity phase imaging confirms 100% cell counting accuracy across variable flow velocities (4-24 frames for a FoV traversal), despite tracking accuracy declining from 97.22% (24 frames for a FoV traversal) to 72.13% (4 frames for a FoV traversal) due to elevated inter-frame displacement. The method eliminates static/non-uniform-moving cell misrecognition while maintaining multimodal compatibility. It is recommended to employ more captures to ensure robust tracking while reduced captures only to ensure accurate counting. This framework provides a robust platform for high-fidelity cell screening in complex hydrodynamic environments.