| In order to study on biological, it is needed to analysis the cells movement. And tracking the cells correctly has played a crucial role in the analysis of cells movement.In the current research on neural stem cells, it is demonstrated that the neural stem cells has a positive effect on the treatment of the cancer and the nervous system diseases because of the neural stem cells are not fully differentiated which can divide to make different types of neurons. So, researching on the differentiation and reproduction law of neural stem cells has become a hotspot in the field of biomedicine.In the analysis of the high density neural stem cells, because the manual tracking method is time-consuming and easy-to-error, it is gradually replaced by the digital image processing method. Therefore, it is required an automatic system which can track high density neural stem cells in the current research.This paper has research on the tracking of the high density cells. And different methods of segmentation are adopted base on the different imaging characteristics of the two testing cell- image sequences. In the part of tracking, according to cell motion characteristics, the cells are classified into two groups- the inactive cells and the active cells.For the Mean Shift algorithm tracking the active cells can easily lead to fail, the topological constraints method can be used to track the active cells. In order to improve the performance of topological constraints algorithm tracking high density cells, area and perimeter factors have been introduced as a new restriction. In addition, the solutions are given to solve the problems of the cells disappearing and the emerging.The presented algorithm is applied in two sequence images of 150 frames. The resulted show that the method proposed in this paper can improve the accuracy rate to 4%～17% and 2%～7% than the Mean Shift and topological constraint respectively. And this algorithm is more effective to track the high density cells. |
PDF Full Text Request