Project Description
Built entirely on the open-source SOFA Framework, DejaVu offers realistic simulations of organ deformable combining surgical images with physics-based modeling and has recenlty been used it to validate deformable registration, 3D reconstruction, and force feedback estimation.
The repository includes five surgical scenes, each with meshes, FEM, textures, and surgical images.
DejaVu is a surgical simulation tool for intra-operative gesture rehearsal based on the SOFA framework. It bridges pre-op simulation and intra-op augmented reality by using real-time physical modeling and intra-operative images to deliver visually accurate, interactive organ simulations for grasping, pulling, and cutting.
It has also been successfully used for generating training datasets for deep learning models and as ground-truth to validate deformable registration, 3D reconstruction, and view synthesis methods, like NeRFs or Gaussian Splats or for force-feedback estimation.
Related publications
DejaVu: Intra-operative Simulation for Surgical Gesture Rehearsal, N. Haouchine et al.
Comments are closed.