Question about pose preprocessing

princeton-vl / DeepV2D

BSD 3-Clause "New" or "Revised" License

653 stars 93 forks source link

Question about pose preprocessing #38

Closed jwh97nn closed 3 years ago

jwh97nn commented 3 years ago

Hi,

I have a question about retrieving the pose data. As referenced below, after the pose is converted from quaternion to matrix, it follows by an inverse operation. Why is this inverse operation necessary?

https://github.com/princeton-vl/DeepV2D/blob/eb362f2f25338faf5adbd7818f1517018bfbc4b5/deepv2d/data_stream/nyuv2.py#L112-L113

Thanks

zachteed commented 3 years ago

The inverse operator converts poses from "world" coordinates to "camera" coordinates. This is necessary because we use camera coordinates for training.