Related Experiment Video
Updated: May 19, 2026
08:59
Morphology-Based Distinction Between Healthy and Pathological Cells Utilizing Fourier Transforms and Self-Organizing Maps
Published on: October 28, 2018
Morphology-based features for adaptive mitosis detection of in vitro stem cell tracking data
1Graduate School for Computing in Medicine and Life Science, University of Lübeck, Ratzeburger Allee 16023538 Lübeck, Germany.
Methods of Information in Medicine
|September 1, 2012
Summary
A new mitosis detector uses maximum likelihood estimation to accurately identify cell division in adult stem cell cultures. This unsupervised method is crucial for autonomous cell farming and enables detailed cell cycle analysis.
Area of Science:
- Cell biology
- Biotechnology
- Regenerative medicine
Background:
- Accurate monitoring of cell division is vital for stem cell production in drug discovery and regenerative medicine.
- Understanding mitotic events aids in reconstructing cell lineages, proliferation curves, and cell cycle analysis.
- Autonomous cell farming requires reliable, unsupervised detection of cell division.
Purpose of the Study:
- To develop and evaluate an unsupervised mitosis detector for adherently growing cell populations.
- To assess the detector's performance across different cell growth phases (lag, log, stationary).
- To compare the proposed method with Support Vector Machines (SVMs).
Main Methods:
- A maximum likelihood (ML) estimator utilizing morphological cell features (area, brightness, length, compactness) was developed.
- The ML approach was designed to adapt to various cell growth phases.
- Performance was evaluated against linear, quadratic, and Gaussian kernel SVMs using the CeTReS benchmark dataset.
Main Results:
- The adaptive ML mitosis detector significantly outperformed non-adaptive methods and linear SVMs.
- The ML approach demonstrated performance comparable to quadratic and Gaussian SVMs.
- The detector reliably distinguishes between mitotic and non-mitotic events.
Conclusions:
- The proposed simple, label-free, adaptive ML mitosis detector is suitable for autonomous cell farming.
- Reliable and unsupervised mitosis detection across all cell growth phases is essential for advanced cell culture applications.
- This method facilitates precise cell cycle analysis and lineage reconstruction.
