Automated tracking approach with ant colonies for different cell population density distribution

Visual tracking of cells has many challenges due to complex situations such as varying cell population densities, intricate motion patterns and sophisticated interactions with other cells. This paper focuses on an efficient and effective ant-based method with working modes updated to track multiple cells over varying densities in the presence of occlusion or clustering situations. To overcome these challenges, the proposed Ant Colony Optimization (ACO) algorithm models two types of ant working modes, namely, cooperation mode and interactive competition mode, whereas the classical ACO algorithms use only one type of mode which performs poor in clutter situations. Moreover, mode update strategies based on ant colony pheromone is used to adjust pheromone field to obtain accurate state vector of cells. Experimental results demonstrate that the proposed method robustly tracks multiple cells in various scenarios. The averaged LSR, LTR and FTR of our method can be only 1.43, 1.71 and 1.37 %, respectively. Our experimental results also show that our tracking method is competitive with state-of-the-art multi-cell tracking methods.