Dielectrophoresis-based classification of cells using multi-target multiple-hypothesis tracking

In this paper we present a novel methodology for classifying cells by using a combination of dielectrophoresis, image tracking and classification algorithms. We use dielectrophoresis to induce unique motion patterns in cells of interest. Motion is extracted via multi-target multiple-hypothesis tracking. Trajectories are then used to classify cells based on a generalized likelihood ratio test. We present results of a simulation study and of our prototype tracking the dielectrophoretic velocities of cells.