✨ About DL_Track_US¶
Mission¶
Our mission is to make deep learning-based muscle architecture analysis accessible to researchers, students, and clinicians working with musculoskeletal ultrasound.
DL_Track_US provides an intuitive graphical interface to segment and analyze fascicles and aponeuroses, allowing cutting-edge research without requiring advanced programming skills.
Goals¶
- Democratize the use of deep learning for ultrasound image and video analysis.
- Provide a user-friendly and customizable software platform.
- Foster open-science practices by making the tool freely available.
- Continuously improve segmentation accuracy through updates and community feedback.
Roadmap¶
✅ Current Version (v0.3.0) includes:
- Automated and manual analysis of images and videos.
- Fascicle, aponeurosis segmentation and muscle parameter extraction.
- Model training module.
- Video preprocessing tools (cropping, resizing, anonymization).
🔜 Future Plans:
- Add intuitive manual labelling tools.
- Combine with feature tracking algorithm, e.g., Lucas-Kanade or kalman filter.
- Finalize fascicle curvature estimation and consideration.
- Extend model architectures (e.g., attention U-Nets, transformers).
- Broaden community contribution possibilities.
Meet the Developers¶
Paul Ritsche¶
Paul is the lead developer and primary coder behind DL_Track_US.
He designed, implemented, and maintains the software in close collaboration with Neil Cronin.
Paul has a background in biomechanics, ultrasound imaging, and software development.
Neil Cronin¶
Neil is the co-developer and main visionary behind DL_Track_US.
His research in biomechanics and muscle architecture guided the core concepts and user-centered design of the tool.
Olivier Seynnes¶
Olivier has been involved with DL_Track_US from the early stages, providing critical input, scientific validation, and feedback during development.
Olivier’s expertise in muscle physiology and musculoskeletal ultrasound ensured DL_Track_US met high scientific standards.