Open YZWYD opened 3 months ago
Hi, thanks for your interest.
A simple method is to use colmap for direct estimation. However, we found that the errors of HyperNeRF and even NeRF-DS are too large. We recommend another more robust method, which you can refer to the process of NeRF-Texture.
Hello, first of all, thank you for open-sourcing such excellent work. I have a small question. If I want to use your code to reconstruct a scene from a video, what other preparations do I need to make besides shooting a video with my phone?