Counting cells from microscopy videos without tracking individual cells

Counting the unique number of cells in a microscopy video (i.e., counting a cell only once while the cell is within the field of view in the video), is required in many biological and pathological studies. Conventionally, cell counting from videos is computed by tracking individual cells. Because tracking cells is non-trivial, these methods are plagued with inaccuracies. In this paper, we engineer a novel and straightforward solution to the problem of unique cell counting by combining frame based cell count with simple pixel motion computation. We estimate the influx and/or the outflux rate of unique cells in a region of interest within the field of view of a microscopy video. The unique count is then obtained by summing the influx and/or the outflux rates. Our proposed framework avoids individual cell tracking altogether; thus, it is capable of overcoming various difficulties associated individual cell tracking. We validate the framework on 11 cell videos of human monocytes. The number of cells is numerous in these videos, yet we obtain a mean counting accuracy of 99%.