Real-Time Dense Deformation Field Estimation of Brain Tissue from Surgical Microscope Video
The proposed project aims to develop and validate a computational method for estimating how brain tissue moves and reacts to surgical tools using only videos captured from a microscope. The student will test both classical and modern deep-learning optical-flow algorithms, adapt them to the specific challenges of microscope imaging, and evaluate how well they can capture small, localized tissue deformations. Controlled experiments on ex-vivo bovine brain specimens will be conducted to collect synchronized videos and ground-truth force measurements, allowing the team to assess the accuracy and reliability of the force-estimation model under realistic surgical conditions. The outcome of this work will be a validated pipeline that can estimate tissue forces without external sensors, enabling more objective, real-time assessment of surgical performance. This project will benefit the participating institutions by advancing low-cost, AI-driven tools for surgical simulation, strengthening collaborative research in medical image analysis, and creating new opportunities for training, knowledge transfer, and future joint publications.
View Full Project DescriptionHoussem Gueziri
École Nationale d'Électronique et des Télécommunications de Sfax
Computer science
Artificial Intelligence; Technology
Université TÉLUQ
Globalink Research Award
