Register now After registration you will be able to apply for this opportunity online.
Uncertainty-aware 3D Mapping
The goal of this project is to enhance the 3D mapping capabilities of a robotic agent by incorporating uncertainty measures into MAP-ADAPT, an incremental mapping pipeline that constructs an adaptive voxel grid from RGB-D input.
The goal of this project is to enhance the 3D mapping capabilities of a robotic agent by incorporating uncertainty measures into MAP-ADAPT [a]. MAP-ADAPT is an incremental mapping pipeline that constructs an adaptive voxel grid from RGB-D input, currently using geometric complexity and semantic information to prioritize high-quality reconstruction of specific regions. This project aims to refine these criteria by integrating local uncertainty measures — arising from estimated camera trajectories, geometry, or semantics — into the decision-making process for reconstruction quality. The primary objective is to achieve a memory- and storage-efficient 3D reconstruction by allocating high detail to complex regions of the scene, while simpler structures (e.g., walls) are represented more coarsely.
[a] Zheng, J., Barath, D., Pollefeys, M. and Armeni, I., 2025. Map-adapt: real-time quality-adaptive semantic 3D maps. In European Conference on Computer Vision (pp. 220-237). Springer, Cham.
The goal of this project is to enhance the 3D mapping capabilities of a robotic agent by incorporating uncertainty measures into MAP-ADAPT [a]. MAP-ADAPT is an incremental mapping pipeline that constructs an adaptive voxel grid from RGB-D input, currently using geometric complexity and semantic information to prioritize high-quality reconstruction of specific regions. This project aims to refine these criteria by integrating local uncertainty measures — arising from estimated camera trajectories, geometry, or semantics — into the decision-making process for reconstruction quality. The primary objective is to achieve a memory- and storage-efficient 3D reconstruction by allocating high detail to complex regions of the scene, while simpler structures (e.g., walls) are represented more coarsely.
[a] Zheng, J., Barath, D., Pollefeys, M. and Armeni, I., 2025. Map-adapt: real-time quality-adaptive semantic 3D maps. In European Conference on Computer Vision (pp. 220-237). Springer, Cham.
