Result description
This innovation integrates high-quality implicit 3D representation into both training data augmentation and test-time optimization, combined with CNN regression for fast initial estimates. Requires only single RGB camera, works in GPS-denied, indoor, outdoor and visually complex environments.
Addressing target audiences and expressing needs
- Collaboration
Collaboration with other research centers and universities to further advance this technology.
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
Continue the research to further advance the technology.
The published results are fully reproducible using the provided source code available at https://github.com/VCL3D/HAL-NeRF

