Long-Term, Accurate Augmented-Reality System for Laparoscopic Videos
Gustavo Puerto and Gian-Luca Mariottini
Augmented-Reality (AR) displays increase surgeon’s visual awareness of high-risk surgical targets (e.g., the location of a tumor) by accurately overlaying pre-operative radiological 3-D model onto the intra-operative laparoscopic video.
However, existing AR systems are not robust to sudden camera motion or prolonged occlusions. These sudden events can cause the loss of anchor points tracked along the video sequence, and thus the loss of the AR display.
Our research team at the ASTRA Robotics lab is working on AR system that are robust to sudden camera changes. We are working in collaboration with the Urology Department at the University of Texas Southwestern to implement an augmented-reality surgical system for partial nephrectomy and prostate cancer scenarios.
Figure 1: Preliminary results: Our system can successfully recover a high percentage of lost tracked features, thus maintaining a long-term augmented display even after prolonged camera occlusions. The last row shows that the pixel reprojection error never goes over 6 pixels.
Figure 2: The video of the experiments.
G. Puerto, G.L. Mariottini, "Real-time, Robust, and Accurate Recovery of Augmented-Reality Display After Occlusions in Laparoscopic Videos", in Medical Image Computing and Computer Assisted Interventions (MICCAI12), Workshop on Augmented Environments for Computer-Assisted Interventions, Nice, France, 2012.
File translated from
version 3.85. On 7 Sep 2012, 16:30.