Deep Learning for Tissue Sensing during Needle Insertions
Betreuer/in: Martin Gromniak
Dekanat/Institut: Institute of Medical Technology and Intelligent Systems
E-Mail: martin.gromniak@tuhh.de
Background: Needle placement is a challenging problem for applications such as biopsy or brachytherapy. Tip force sensing can provide valuable feedback for needle navigation inside the tissue. For this purpose, fiber-optical sensors can be directly integrated into the needle tip. We employ optical coherence tomography (OCT) to both image the tissue in front of the needle and estimate forces on the needle tip. As the intepretation of the OCT data is challenging we use neural networks to process it in real-time.
Tasks:
- Literature Research on spatio-temporal deep learning models
- Familiarize with the data for tip force estimation and tissue imaging
- Implement a training pipeline
- Evaluate different training approaches and network architectures with respect to accuracy, robustness and inference speed.
This topic can be scaled to a MSc, BSc or project thesis.
Requirements: Good programming skills, prior practical knowledge in machine learning, ability to work independently
