Detection And Tracking Of Human Sperm Cells

In the field of medicine and bio-pharmaceutical, cells motion analysis has played a crucial role in the disease control and efficacy testing. The first step of cells motion analysis is cell tracking, achieving all kinds of data in the process of cell movement as the research materials.In this paper, sperm cells are used as the research object, we propose a new method to automatically extract the motion of objects in image sequences. The first step use the background subtraction to detect moving sperm cells, but there are some limitations to select an appropriate threshold value since the output accuracy is strongly dependent on the selected threshold value. To eliminate this dependency, we propose an improved non-linear diffusion filtering in background subtraction. By applying the non-linear diffusion algorithm, we determine a more proper threshold value and can obtain an accurate motion mask which isn’t contaminated by noise.Then, we improve the conditions of weight functions and the standard of belief-propagation flows of graph theory. First we use an efficient method to map the cells in the single frame become completely undirected graph. To simplify the processing of cells tracking to vertex matching between the two graphs.The results confirm that the detection algorithm identify progressive sperms very well. And the improved method of graph theory hence the accuracy rate was 92%. It is proved that the sperm cells tracking algorithm in this paper is greatly improved in comparison with the mean shift tracking method, the standard particle filter, combination of Kalman and particle filter. Our approach performs with comparable computational efficiency.