1. Augmented Reality(AR) Based Tooth Preparation Training System
Tooth preparation is a surgical procedure in which dentists utilize a drilling tool to remove enamel and dentin, making the tooth the desired shape for the subsequent placement of a crown or bridge. In practice, tooth preparation involves four steps: creating the occlusal surfaces, reducing the buccal surfaces, cutting half of the lingual surfaces, and smoothing these surfaces. The standard depth requirement for occlusal and buccal reduction should fall within a range of approximately 1 mm to 1.5 mm. Excessive drilling poses the risk of damaging the pulp and causing inflammation, while inadequate drilling may result in an improper fit for the crown size.
To accurately complete these steps of tooth preparation, dentists must possess highly delicate surgical skills and engage in long-term practice.The objective of this project is to develop a training system for tooth drilling while providing real-time force feedback by using haptic device and visualizing digital tooth on AR Headset HoloLens2. This implementation aims to make surgeons familiarize the tooth preparation steps for the installation of a custom-made dental crown.
2. Realtime Robust Shape Estimation of Deformable Linear Object
Realtime shape estimation of continuum objects and manipulators is essential for developing accurate planning and control paradigms. The existing methods that create dense point clouds from camera images, and/or use distinguishable markers on a deformable body have limitations in realtime tracking of large continuum objects/manipulators. The physical occlusion of markers can often compromise accurate shape estimation. We propose a robust method to estimate the shape of linear deformable objects in realtime using scattered and unordered key points. By utilizing a robust probability-based labeling algorithm, our approach identifies the true order of the detected key points and then reconstructs the shape using piecewise spline interpolation. The approach only relies on knowing the number of the key points and the interval between two neighboring points. We demonstrate the robustness of the method when key points are partially occluded. The proposed method is also integrated into a simulation in Unity for tracking the shape of a cable with a length of 1m and a radius of 5mm. The simulation results show that our proposed approach achieves an average length error of 1.07% over the continuum’s centerline and an average cross-section error of 2.11mm. The real-world experiments of tracking and estimating a heavy-load cable prove that the proposed approach is robust under occlusion and complex entanglement scenarios. View More Details
3. Cardiac Ultrasound Image Segmentation and Stroke Volume estimation
Cardiac ultrasound imaging has been widely used around the world for providing diagnosis of heart diseases and intraoperative support. The information of certain regions in cardiac ultrasound images reflects patients’ heart health conditions. Manually delineating the region of interest in cardiac ultrasound image is not efficient and not stable in practice. This project applied two U-Net neural networks to automatically segment left ventricle endocardium for 2-chamber view and 4-chamber view in cardiac ultrasound images of 250 patients. Based on segmented ultrasound images from different views in a sequence, the 3D shape of the left ventricle endocardium can be constructed to estimate the stroke volume.
