facebookresearch / facebookresearch/OrienterNet
About roll angle rectify after image flipping
- Dominant language
- Python
- Stars
- 577
- Forks
- 66
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
I noticed that in [line 159](https://github.com/facebookresearch/OrienterNet/blob/c222d08c9c0a084d79a415cc5c43e71173d2ecc4/maploc/data/dataset.py#L159), you employed random flipping to augment the query and map images, and later, in [line 164](https://github.com/facebookresearch/OrienterNet/blob/c222d08c9c0a084d79a415cc5c43e71173d2ecc4/maploc/data/dataset.py#L164), the query image are rectified by setting the roll and pitch angles to zero. However, I'm wondering if a negative sign should be applied to the [roll](https://github.com/facebookresearch/OrienterNet/blob/c222d08c9c0a084d79a415cc5c43e71173d2ecc4/maploc/data/dataset.py#L214) angle during the rectification process after the images have been flipped.
I'm quite new to this area, thank you for your attention to this detail!
Contributor guide
Assessment
This issue has not been assessed yet.