An Ant Colony Inspired Multi-Bernoulli Filter for Cell Tracking in Time-Lapse Microscopy Sequences

Area of Science:

• Cell biology
• Computational biology
• Biophysics

Background:

• Accurate cell tracking in time-lapse microscopy is crucial for understanding dynamic cellular processes.
• Existing methods struggle with complex behaviors like mitosis, morphological changes, and erratic cell movements.

Purpose of the Study:

• To develop a robust and automatic cell tracking method for analyzing dynamic cell behaviors in microscopy sequences.
• To estimate individual cell states and delineate cell lineage trees.

Main Methods:

• Proposed a novel ant colony-inspired multi-Bernoulli filter for tracking cell collections.
• Modeled each ant colony as independent, using pheromone fields and heuristic information for density approximation.
• Implemented a dual prediction mechanism for ant colonies and pheromone fields to guide tracking between frames.

Main Results:

• Demonstrated effective tracking of cells with varying densities, frequent mitosis, and uneven motion.
• The proposed algorithm significantly outperformed recently reported cell tracking approaches.
• Successfully estimated cell states and delineated lineage trees in challenging datasets.

Conclusions:

• The ant colony-inspired multi-Bernoulli filter provides a reliable and automatic solution for complex cell tracking tasks.
• This approach enhances the analysis of dynamic cellular behaviors in microscopy.
• The method shows significant potential for advancing cell biology research.