CELL TRACKING USING PARTICLE FILTERS WITH IMPLICIT CONVEX SHAPE MODEL IN 4D CONFOCAL MICROSCOPY IMAGES.

Area of Science:

• Computational Biology
• Biophysics
• Developmental Biology

Background:

• Bayesian frameworks, particularly particle filters, are vital for tracking algorithms.
• Particle filters utilize stochastic motion and observation models for state evolution and measurement integration.
• Tracking cell lineage is crucial in developmental biology, with shape propagation offering potential benefits.

Purpose of the Study:

• To develop an advanced particle filter capable of tracking cell lineage by incorporating cell shape information.
• To introduce a novel method for approximating arbitrary cell shapes using implicit convex functions.
• To validate the efficacy of the proposed shape-aware particle filter in tracking cell lineages in vivo.

Main Methods:

• Approximation of arbitrary cell shapes using a novel implicit convex function.
• Definition of the particle filter's importance sampling step based on the cost of fitting the implicit convex shape model to observations.
• Application of the algorithm to track cell lineage during nonmitotic stages.

Main Results:

• The developed technique successfully tracks cell lineage by integrating cell shape models into the particle filter.
• The implicit convex function effectively approximates complex cell shapes.
• The algorithm demonstrated capability in tracking the lineage of retinal and lens cells in zebrafish embryos.

Conclusions:

• Integrating cell shape models into particle filters enhances cell lineage tracking accuracy.
• The novel implicit convex shape function provides an effective way to represent and propagate cell shapes.
• This method offers a robust approach for tracking cell lineages in developmental studies, particularly in nonmitotic phases.