FastTrack

FastTrack extracts trajectories of deformable objects from two-dimensional movies for quantitative analysis of dynamic biological and physical systems.

Key Features:

• Trajectory extraction: Extracts object trajectories from two-dimensional movies across various length scales, including deformable objects such as cells and animals.
• Optimized performance: Handles a dynamic number of deformable objects within a region of interest and runs orders of magnitude faster than state-of-the-art tracking algorithms while maintaining comparable accuracy.
• Trackability metric ("probability of incursions"): Computes the "probability of incursions" to assess a movie's trackability without requiring ground truth, with resilience to minor errors and computable via ad hoc tracking.
• Automated parameter optimization: Uses an iterative algorithm to determine nearly-optimized tracking parameters with minimal human input, enabling exploration of parameter space and optimization of the number of swaps for batches of similar movies.
• Open-source integration: Distributed under GNU GPLv3 with a core that can be integrated into C++/Qt projects.

Scientific Applications:

• Cell tracking and motility: Quantifies cell trajectories and cell motility from two-dimensional movies.
• Animal tracking and behavior: Extracts animal trajectories for behavioral analysis.
• Fluid dynamics: Tracks deformable objects within fluid flow recordings for dynamic analysis.
• Material science: Analyzes dynamics of deformable components in material systems from video recordings.

Methodology:

Extracts trajectories from two-dimensional movies; computes the "probability of incursions" using ad hoc tracking to assess trackability; and applies an iterative algorithm to determine nearly-optimized tracking parameters by optimizing the number of swaps across batches of similar movies.