Refactor Bounding Box Code

[OrangePear13]

Refactor Bounding Box Code

Problem

The Bounding Box code in vision/common/bounding_box.py needs to be refactored to reduce complexity and remove unnecessary features.

Solution

The refactored bounding_box.py should:

• not have setter and getter functions
  • unnecessary for our use case
• not have an ObjectType
  • no longer needed (also remove the Enum)
• not have an attributes dictionary
  • largely unnecessary and should be implemented seperately if needed
• Store coordinates as a top left corner position, width, and height
  • NOTE: Tentative, further discussion needed

[J-The-Fox]

Does anything need to be done with the tlwh_to_vertices function as well or can that be left how it is?

[OrangePear13]

I think it should be fine as is, at least for now.
It might become redundant if we convert the bounding box representation to the format of storing one corner coordinate along with a width and length. But, I still haven't decided if that would be a good idea, as I'm not yet sure if we will need support for rotated boxes, TBD.