An accurate different density distributions cell parameter estimate algorithm based on ant colony optimization

This paper focuses on a new ant-based algorithm for accurately tracking multiple cells of different population density distributions in each frame. Two types of ant working modes are modeled and correspond to two events, namely, interactive competition mode and cooperation mode. In a high cell density region, adjacent ant colonies with interactive competition mode encourages ant colonies to work between cooperation and competition, whereas in a low density region, ant colonies with the cooperation mode introduces simple pure cooperative mechanism in order to search for cells quickly. To further obtain accurate individual state, mode update strategy based on ant colonies interaction mechanism is utilized to adjust pheromone field. Finally, we compared the experiment results on real-world data with those reported in the most current literature. Simulation results demonstrate that our algorithm can automatically and accurately track numerous cells in various scenarios, and is competitive with state-of-the-art multi-cell tracking methods.