Mehak Gupta

Semester Work
Supervisors: Janine Thoma, Dr. Ajad Chhatkuli, and Prof. Luc Van Gool

Improving Pose-estimation accuracy in the context of Augmented Reality

To ease the dental scanning procedure using intra-oral dental scanners, Thoma et al. [1] have developed an AR application for the overlay of the 3D computer teeth model on the patient's teeth in real time. In this semester project, we investigate the possibility of improvement and verification of the pose estimation in order to improve augmentation of the scanned 3D teeth model. Firstly, we use blob markers on the intra-oral scanner for improvement of pose estimation. We evaluate the changes by reconstructing and using a reference model. Secondly, we verify the existing pose estimation technique by Mutual Information. We show with quantitative results that blob markers cannot further improve the results. On the other hand, we show with qualitative results that the refinement method based on Mutual Information can further improve the estimated camera pose. References: [1] J. Thoma, M. Havlena, S. Stalder and L. V. Gool, "[POSTER] Augmented Reality for User-Friendly Intra-Oral Scanning," 2017 IEEE International Symposium on Mixed and Augmented Reality (ISMAR-Adjunct), Nantes, 2017, pp. 97-102.