Interdisciplinary Institute in Artificial Intelligence

DESCRIPTION

The chair focuses on the use of artificial intelligence in the context of developing computer-assisted medical and surgical procedures. This includes AI-based image processing and calibration, AI-based simulation, and extracting knowledge from procedure records to model procedure quality. Context-aware assistance and more autonomous assistive devices (including robots) are also addressed within the scope of the chair.
The chair is run in close collaboration with clinical teams and industry partners.

ACTIVITIES

The work already undertaken focuses on the use of deep learning for the automatic calibration of interventional CBCT imaging systems. Deep learning is also being studied for image processing and fusion to enable real-time guidance of diagnostic or treatment procedures such as prostate biopsies and neurosurgery. In both of these cases, ultrasound imaging plays a predominant role during the procedure and poses significant challenges in terms of image processing. In addition, we are developing new robotic approaches based on continuous robots, with the medium-term goal of endoluminal application.

CHAIR EVENTS

Organizing AI-related activities at the national level with CAMI partners. Working group on AI for Health at TIMC and CIC/CHU Grenoble Alpes.

SELECTED LIST OF PUBLICATIONS 

• Dupuy, T., Beitone, C., Troccaz, J., & Voros, S. (2021). 2D/3D Deep Registration for Real-Time Prostate Biopsy Navigation. Accepted for publication in SPIE Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling. February 2021.
• Younes, H., Troccaz, J., & Voros, S. (2021). Machine learning and registration for automatic seed localization in 3D ultrasound images for prostate brachytherapy. Medical Physics.
• Derathé A, Reche F, Jannin P, Moreau-Gaudry A, Gibaud B, Voros S. Explaining a model predicting the quality of surgical practice: an initial presentation to and review by clinical experts. Int J Comput Assist Radiol Surg. June 18, 2021. doi: 10.1007/s11548-021-02422-0. Published online ahead of print. PMID: 34143373.
• Carton, F. X., Chabanas, M., Munkvold, B. K., Reinertsen, I., & Noble, J. H. (2020). Automatic segmentation of brain tumors in intraoperative ultrasound images using 3D U-Net. In SPIE Medical Imaging 2020: Image-Guided Procedures, Robotic Interventions, and Modeling (Vol. 11315, p. 113150S), Houston, March 2020.
• Lapouge, G., Younes, H., Poignet, P., Voros, S., & Troccaz, J. (2019). Needle Segmentation in 3D Ultrasound Volumes Based on Machine Learning for Needle Steering. Hamlyn Symposium on Medical Robotics, London, June 2019.