search

city-super / BungeeNeRF

[ECCV22] BungeeNeRF: Progressive Neural Radiance Field for Extreme Multi-scale Scene Rendering
https://city-super.github.io/citynerf
MIT License
565 stars 61 forks source link

How to run this code on the custom data? #4

Closed DapengFeng closed 2 years ago

kam1107 commented 2 years ago

If you already have images and poses, you need to partition your data to different scale and record where you split it. E.g. in our Google earth datasets, the camera gradually moves far away. So in pose_enu.json, the entry "scale_split": [325, 195, 65, 0] means: stage0(furthest scale) ranges from image index 325 until the end; stage1(a closer scale) ranges from image index 195 until 325 etc.

bycloudai commented 2 years ago

Hi @kam1107 what did you guys used to estimate the poses? thank you