A novel multiple particle tracking algorithm for noisy in vivo data by minimal path optimization within the spatio-temporal volume

Automated tracking of fluorescent particles in living cells is vital for subcellular stoichoimetry analysis. Here, a new automatic tracking algorithm is described to track multiple particles, based on minimal path optimization. After linking feature points frame-by-frame, spatio-temporal data from time-lapse microscopy are combined together to construct a transformed 3D volume. The trajectories are then generated from the minimal energy path as defined by the solution of the time-dependent partial differential equation using a gray weighted distance transform dynamic programming method. Results from simulated and experimental data demonstrate that our novel automatic method gives sub-pixel accuracy even for very noisy images.