Uncertainty-aware asynchronous scattered motion interpolation using Gaussian process regression.

We address the problem of interpolating randomly non-uniformly spatiotemporally scattered uncertain motion measurements, which arises in the context of soft tissue motion estimation. Soft tissue motion estimation is of great interest in the field of image-guided soft-tissue intervention and surgery navigation, because it enables the registration of pre-interventional/pre-operative navigation information on deformable soft-tissue organs. To formally define the measurements as spatiotemporally scattered motion signal samples, we propose a novel motion field representation. To perform the interpolation of the motion measurements in an uncertainty-aware optimal unbiased fashion, we devise a novel Gaussian process (GP) regression model with a non-constant-mean prior and an anisotropic covariance function and show through an extensive evaluation that it outperforms the state-of-the-art GP models that have been deployed previously for similar tasks. The employment of GP regression enables the quantification of uncertainty in the interpolation result, which would allow the amount of uncertainty present in the registered navigation information governing the decisions of the surgeon or intervention specialist to be conveyed.

Projector-based surgeon-computer interaction on deformable surfaces.

PURPOSE: Providing intuitive and easy to operate interaction for medical augmented reality is essential for use in the operating room. Commonly, intra-operative navigation information is displayed on an installed monitor, requiring the operating surgeon to change focus from the monitor to the surgical site and vice versa during navigation. Projector-based augmented reality has the potential to alleviate this problem. The aim of our work is to use a projector for visualization and to provide intuitive means for direct interaction with the projected information. METHODS: A consumer-grade projector is used to visualize preoperatively defined surgical planning data. The projection of the virtual information is possible on any deformable surface, and the surgeon can interact with the presented virtual information. A Microsoft Kinect camera is used to capture both the surgeon interactions and the deformations of the surface over time. After calibration of projector and Kinect camera, the fingertips are localized automatically. A point cloud surface representation is used to determine the surgeon interaction with the projected virtual information. Interaction is detected by estimating the proximity of the surgeon's fingertips to the interaction zone and applying projector-Kinect calibration information. Interaction is performed using multi-touch gestures. RESULTS: In our experimental surgical scenario, the surgeon stands in front of the Microsoft Kinect camera, while relevant medical information is projected on the interaction zone. A hand wave gesture initiates the tracking of the hand. The user can then interact with the projected virtual information according to the defined multi-touch-based gestures. Thus, all information such as preoperative planning data is provided to the surgeon and his/her team intra-operatively in a familiar context. CONCLUSION: We enabled the projection of the virtual information on an arbitrarily shaped surface and used a Microsoft Kinect camera to capture the interaction zone and the surgeon's actions. The system eliminates the need for the surgeon to alternately view the surgical site and the monitor. The system eliminates unnecessary distractions and may enhance the surgeon's performance.