In the pipeline, augmentation such as flip, rotation changes input image.
but these augmentations are not currently being applied to the occupancy GT.
Isn't it logically right to stop applying that augmentations?
The augmentation strategy applied in image view space will not change the spatial distribution of the features in BEV space. You can refer to BEVDet's paper
In the pipeline, augmentation such as flip, rotation changes input image. but these augmentations are not currently being applied to the occupancy GT. Isn't it logically right to stop applying that augmentations?