3D Reconstruction & Neural Rendering

On-the-fly 3D reconstruction has become a fundamental task in numerous augmented and virtual reality (AR/VR) applications that involve fast-changing environments, e.g., virtual room planner, VR painting, metaverse 3D asset creation, and virtual telepresence. 

Our work focuses on empowering edge devices with real-time, energy-efficient 3D reconstruction platforms.

Related Publication:

1. Li S, Li C, Zhu W, et al. Instant-3D: Instant Neural Radiance Field Training Towards On-Device AR/VR 3D Reconstruction[C]//Proceedings of the 50th Annual International Symposium on Computer Architecture. 2023: 1-13.

2. Fu Y, Ye Z, Yuan J, et al. Gen-NeRF: Efficient and Generalizable Neural Radiance Fields via Algorithm-Hardware Co-Design[C]//Proceedings of the 50th Annual International Symposium on Computer Architecture. 2023: 1-12.

3. Zhao Y K, Wu S, Zhang J, et al. Instant-NeRF: Instant On-Device Neural Radiance Field Training via Algorithm-Accelerator Co-Designed Near-Memory Processing[C]//2023 60th ACM/IEEE Design Automation Conference (DAC). IEEE, 2023: 1-6.

4. Li C, Li S, Zhao Y, et al. RT-NeRF: Real-Time On-Device Neural Radiance Fields Towards Immersive AR/VR Rendering[C]//Proceedings of the 41st IEEE/ACM International Conference on Computer-Aided Design. 2022: 1-9.