DeepTrack2
DeepTrack2 is a modular Python library for generating, manipulating, and analyzing image data pipelines for machine learning and experimental imaging.
What is DeepTrack2?
DeepTrack2 is a modular Python library for generating, manipulating, and analyzing image data pipelines for machine learning and experimental imaging.
Overview of DeepTrack2
What it is
DeepTrack2 is a modular Python library for generating, manipulating, and analyzing image data pipelines for machine learning and digital microscopy. It is designed to support experimental imaging workflows and requires at least Python 3.10. Version 2.0 and subsequent do not support TensorFlow; legacy version 1.7 is available for TensorFlow support.
Key features
- Modular Python library for building image data pipelines.
- Supports simulation of various microscopy modalities (e.g., brightfield, fluorescence, holography, TEM, SMLM).
- Includes tools for particle tracking, aberration characterization, and object detection.
- Provides integration with PyTorch for differentiable optimization and gradient-based fitting.
- Offers LodeSTAR for label-free particle detection and MAGIK for graph-neural-network-based particle linking and trace characterization.
- Includes a comprehensive set of tutorials and examples for beginners and advanced users.
Use cases
- Simulating and analyzing microscopy images for training machine learning models.
- Tracking single or multiple particles in experimental videos.
- Characterizing optical aberrations and optimizing phase masks.
- Segmenting cellular structures (e.g., membranes, mitochondria) in electron microscopy images.
- Detecting and counting cells or particles in fluorescence, brightfield, or holographic images.
Getting started
To install DeepTrack2, open a terminal or command prompt and run:bashpip install deeptrackorbashpython -m pip install deeptrackFor detailed getting started guides, see the repository README.
License and pricing
DeepTrack2 is open source under the MIT license.
Based on the project's public README and repository metadata.