Related Experiment Video
Updated: Aug 11, 2025
3D Orbital Tracking in a Modified Two-photon Microscope: An Application to the Tracking of Intracellular Vesicles
Published on: October 1, 2014
High precision tracking analysis of cell position and motion fields using 3D U-net network models
Li-Xin Yuan1, Hong-Mei Xu1, Zi-Yu Zhang1
1International Research Centre for Nano Handling and Manufacturing of China, ChangchunUniversity of Science and Technology, Changchun, 130022, China; Ministry of Education Key Laboratory for Cross-Scale Micro and Nano Manufacturing, Changchun University of Science and Technology, Changchun, 130022, China.
This study introduces a new cell tracking model, Cell Position and Motion Fields (CPMF), for analyzing cell behavior. The model accurately tracks cell migration and division without manual annotation, advancing biological research.
Area of Science:
- Cell biology
- Bioengineering
- Medical imaging
Background:
- Quantitative analysis of cellular states is crucial for understanding biological mechanisms and drug actions.
- Accurate tracking of cell migration and division is essential for studying cell cycle changes and cellular processes.
Purpose of the Study:
- To develop a novel engineering model for high-accuracy cell tracking.
- To enable automated analysis of cell migration, division, and movement within the field of view.
Main Methods:
- Proposed a Cell Position and Motion Fields (CPMF) model.
- Utilized a U-Net network model with 3D Convolutional Neural Networks (CNNs).
- Combined detection and correlation techniques with self-editing training samples.
Main Results:
- Achieved an average cell coordinate detection accuracy of 98.38%.
- Reached an average cell tracking accuracy of 98.70%.
- Demonstrated high-accuracy tracking of cell migration, division, and entry/exit in all directions.
Conclusions:
- The CPMF model offers a robust and accurate solution for cell tracking.
- Automated, high-accuracy cell tracking facilitates deeper insights into cellular dynamics and drug effects.
- The model's ability to train without manual annotation enhances its practical applicability.
