ARAMIS: Augmented Reality Assistance for Minimally Invasive Surgery Using a Head-Mounted Display
We propose ARAMIS, a solution to provide real-time “x-ray see-through vision” of a patient’s internal structure to the surgeon, via an optical see-through head-mounted display (OST-HMD), in minimally invasive laparoscopic surgery. ARAMIS takes input imaging from a binocular endoscope, reconstructs a dense point cloud with a GPU-accelerated semi-global matching algorithm on a per-frame basis, and then wirelessly streams the point cloud to an untethered OST-HMD (currently, Microsoft HoloLens) for visualization. The OST-HMD localizes the endoscope distal tip by fusing fiducial-based tracking and self-localization. The point cloud is rendered on the OST-HMD with a custom shader supporting our data-efficient point cloud representation. ARAMIS is able to visualize the reconstructed point cloud (184k points) at 41.27 Hz with an end-to-end latency of 178.3 ms. A user study with 25 subjects, including 2 experienced users, compared ARAMIS to conventional laparoscopy during a peg transfer task on a deformable phantom. Results showed no significant difference in task completion time, but users generally preferred ARAMIS and reported improved intuitiveness, hand-eye coordination and depth perception. Inexperienced users showed a stronger preference for ARAMIS and achieved higher task success rates with the system, whereas the two experienced users indicated a slight preference for ARAMIS and succeeded in all tasks with and without assistance.
KeywordsAugmented Reality Minimally invasive surgery Laparoscopic surgery Head-mounted display Microsoft hololens
Supplementary material 1 (mp4 40532 KB)
- 9.Qian, L., Azimi, E., Kazanzides, P., Navab, N.: Comprehensive tracker based display calibration for holographic optical see-through head-mounted display. arXiv preprint arXiv:1703.05834 (2017)
- 12.Wang, J., Qian, L., Azimi, E., Kazanzides, P.: Prioritization and static error compensation for multi-camera collaborative tracking in augmented reality. In: IEEE Virtual Reality (VR), pp. 335–336. IEEE (2017)Google Scholar