search

caiyuanhao1998 / SAX-NeRF

"Structure-Aware Sparse-View X-ray 3D Reconstruction" (CVPR 2024)
https://arxiv.org/abs/2311.10959
MIT License
465 stars 19 forks source link

How to train using 2D RGB X-ray images #13

Closed jaykshirsagar05 closed 3 months ago

jaykshirsagar05 commented 3 months ago

Is it possible to use this framework on 2D RGB X-ray images instead of CT scans?

Here are my data specifications and I want to know if it is sensible and possible to train it using your designs:

  1. Collection of 2D rgb xray images
  2. Lacks camera parameters
  3. Dynamic in nature
caiyuanhao1998 commented 3 months ago

Hi, thanks for your interest. Actually, our method is trained on 2D X-ray images instead of CT scans.

If you want to train on your collected 2D RGB x-ray images, the first step is to compute the camera poses for your images. There are two open-source tools for you to use:

DusT3R : https://github.com/naver/dust3r

COLMAP (SfM) : https://github.com/colmap/colmap

Wish this information could help you