Induced Pluripotent Stem Cells Recognition and Localization
Induced pluripotent stem cells (iPSCs) are pluripotent stem cells formed by artificial induced differentiation of somatic cells into Stem cells, which can differentiate into different kinds of cells with different functions. However, the process of inducing stem cell culture and preparation has always been the focus of research, especially the induction and purification of stem cells. It relies mainly on professionals for visual recognition and manual operation, which is not only inefficient, but also prone to cell contamination. Here we are reporting a method based on deep learning to contactlessly identify and locate induced stem cells, and combine high-precision robots to automatically classify and pick cells. First, the trained Faster RCNN model is used to identify the cell image, and the coordinates and categories of the cell region are obtained. Secondly, the trained Alexnet network is used to sequentially traverse the cell regions and classify to obtain the precise region of the cells in the regions. Thirdly, sequentially perform the above operations on all the pictures of the cell culture dish area, and then combining the obtained picture in order. Finally, the fusion of the cell regions is performed on the result of the splicing, and then the coordinates and categories of cell in the culture dish region are obtained. Experiments show that the recognition accuracy of iPSCs is 99%. It can provide a basis for automated picking of iPSCs.