Innovative Al Techniques for Accurate Analysis of Cell Morphology and Viability

Accurate segmentation of cell boundaries is the first and most critical step underlying the quantification of cell morphology. Label-free imaging enables cell behavior to be monitored in a non-perturbing environment, yet images can be challenging to analyze. Cells in culture have widely varying and fluid morphologies with certain types of cells appearing almost transparent; they grow from sparsely populated to highly confluent. In such a dynamic system, continuously and accurately detecting the boundaries between each cell presents a challenge for traditional computer vision approaches. Here we demonstrate the advantages of using a widely trained neural network for segmentation of different cellular models and using the resulting morphological descriptors to classify cells using a machine learning-based classifier.

The use of label-free imaging for long-term cell monitoring is a powerful approach for examining live cells, providing longitudinal insight into cell behavior, function, and response to external treatment or stimulus.

Key Benefits:

• Improved segmentation enables more accurate morphology classification.
• Microglia viability and activation can be continuously monitored with label-free analysis.
• Non-toxic drug effects can be quantified.
• Transient morphology indicates ferroptotic cell death.