Biological cell motion tracking in dielectrophoresis (DEP) levitation feedback control system

In this paper, a novel hierarchical adaptive merge/split structured mesh (HAMSM) approach that is used in biological cell tracking is presented. This approach takes advantage of the motion of a single biological cell to extract it. After generating a moving object region in a current frame, the best match is determined in the next frame by performing image correlation matching, computed by convolving the object´s intensity template over candidate regions in the new image. This method is not influenced by the blurred edges of the biological cells. It outperforms the intensity-based algorithms. This approach needs no manual initialization. In this paper it is demonstrated on a biological cell sequence and is compared with the frame difference (FD) algorithm and the dense motion fields algorithm.