| Type | Semester project |
| Split | 40% theory, 30% software, 30% experiment |
| Knowledge | Python, deep learning; Pytorch, MuJoCo and human behaviour modeling a plus |
| Subjects | Software development, data analysis |
| Supervision | Luyin Hu, Soheil Gholami |
Surgical training is moving toward objective skill assessment to reduce subjectivity and bias. This project analyzes hand motion data from a microsurgical anastomosis task, integrating state-of-the-art hand–object interaction techniques with simulation-based analysis of anatomical force and stiffness properties to reveal the principles that differentiate surgical performance and expertise.
Approach
- Review methods for hand–object pose reconstruction, pattern recognition, and biomechanical simulation.
- Preprocess and extract features from hand motion data; assess the added value of force measurements.
- Implement and compare algorithms to identify performance-related patterns.
Expectation
- Strong self-motivation and interest in deep technical exploration.
- Well-documented, reproducible code for data processing and modeling.
- Clear and concise report summarizing methodology, results, and insights.