Stable marriage algorithm for tracking intracellular objects

Development of automatic multiple intracellular-objects tracking methods is one of the significant challenges in Bioimage-Informatics. The challenge becomes more difficult in case the tracking targets have the same shape and appearance. In order to obtain stable results under that condition, we propose a tracking method based on global optimization. Particularly, we first detect tracking targets by our proposed detection method. Then, we formulate the multiple object tracking problem as a combinatorial optimization problem over a pair of consecutive frames. Finally, we solve the problem by the stable marriage algorithm. In this paper, we describe our proposed detection and tracking methods.