[RFC] Implement transforms primitives for Bounding Boxes

[vfdev-5]

Related tracker issue: #5754

Functional transforms for bboxes

The idea is to implement the following low-level functional transforms for bounding boxes:
https://github.com/pytorch/vision/blob/main/torchvision/prototype/transforms/functional/_geometry.py

• affine_bounding_box (@vfdev-5) [proto] Added functional affine_bounding_box op #5597
• rotate_bounding_box (@vfdev-5) [proto] Added functional rotate_bounding_box op #5638
• perspective_bounding_box (@vfdev-5) [proto] Added functional perspective_bounding_box/segmentation_mask ops #5888
• center_crop_bounding_box (@vfdev-5) [proto] Added center_crop_bounding_box functional op #5972
• crop_bounding_box (@vfdev-5) [proto] Added functional crop_bounding_box op #5781
• resized_crop_bounding_box (@vfdev-5) [proto] Added resized_crop_bounding_box op #5853
• pad_bounding_box (@pmeier) port RandomZoomOut from detection references to prototype transforms #5551, add tests for F.pad_bounding_box #6038
• vertical_flip_bounding_box refactor: port RandomVerticalFlip to prototype API (#5524) #5633

How to test the implementation ?

Put tests into test/test_prototype_transforms_functional.py:

• correctness test
• eager vs scripted
• scriptable
• check on cpu and cuda

cc @vfdev-5 @datumbox @bjuncek