Deep manifold learning of cell morphology and motion
Organization:
Funded by: | EEMCS Faculty |
PhD: |  |
Supervisor: | chair MIA:  |
Collaboration: |      |
Description:
Information about the characteristics of cells is important. For instance, understanding the shape properties of a cell over time, may give insight into the distinction between good and bad cells. However, cell populations, such as populations of sperm cells and protoplasts, are often analyzed by looking at average values of a population. Only considering average quantities is not ideal as a cell population is not homogeneous; it consists of different individual cells with different responses to stimuli. As a consequence, it is beneficial to study the properties of single cells.
Microfluidic technologies are a perfect tool to analyze the properties of individual cells. Using such technologies, videos can be obtained of the moving cells. These images contain a plethora of information about the cells. This project is aimed at using Geometric Deep Learning to generate more knowledge about the heterogeneity of cells in a population. More precisely, a latent encoding is created that encodes an informative, understandable, and discriminable feature representation of cells. The challenge is to extract informative cell morphology like shape, structure, form or size, and informative cell dynamics like velocity, acceleration, rotation, deformation or grows from dynamic imaging data and combine it into a unified higher-order feature representation. The resulting Deep Learning based manifolds can then be explored visually (explainable AI) and used for important classification questions like: Which is the optimal sperm cell to choose? What is a perfectly growing protoplast?
Output:
Pictures:
"EEMCS Theme call SEARCH Team"