search

facebookresearch / co3d

Tooling for the Common Objects In 3D dataset.
Other
963 stars 75 forks source link

Format of camera poses listed in frame_annotations.jgz #84

Closed justachetan closed 9 months ago

justachetan commented 9 months ago

Hi! Are the camera poses (rotation and translation vector) listed in frame_annotations.jgz file in each category sub-folder in the world-to-camera format or the camera-to-world format?

Any help on this would be much appreciated!

shapovalov commented 9 months ago

The cameras are in Pytorch3D format, and if you use the provided data loaders, they will be loaded correctly. In case you use different frameworks, it should be: x_cam = x_world @ R + T (and then decalibrate with intrinsics if needed). More info: https://pytorch3d.org/docs/cameras