search

VITA-Group / SinNeRF

[ECCV 2022] "SinNeRF: Training Neural Radiance Fields on Complex Scenes from a Single Image", Dejia Xu, Yifan Jiang, Peihao Wang, Zhiwen Fan, Humphrey Shi, Zhangyang Wang
MIT License
329 stars 26 forks source link

Question about the rectified DTU dataset. #30

Closed pansanity666 closed 1 year ago

pansanity666 commented 1 year ago

Hi, I have found that your used DTU seems to be different from the commonly used one by pixelNeRF, DietNeRF, etc. Specifically, the images seems to be closer than the standard one. e.g.: standard scan8 cam0: image

yours: image

May I ask how did you preprocess the DTU dataset? Did you crop the image? This is important for me to perform fair comparisons.

Best,

Ir1d commented 1 year ago

HI @pansanity666 , If I remember correctly, I used https://github.com/xy-guo/MVSNet_pytorch to obtain depth on the DTU dataset. So I assume my DTU dataset was downloaded from their codebase as well. https://github.com/xy-guo/MVSNet_pytorch#training I didn't do any further processing for the images.

Hope this helps!