Recognition with learning method

TODO:

• Check if reco ROI affects the input image
  -> If yes, "reset" ROI to be the full image and just discard detections outside of it
• adjust pet files when !285 (merged) is finished
• fix weird detection bug
• add tool tips to parameters
• check if number/name of classes can be read out in the output
• add support for YOLOv8
• add to trajectory instead of creating a new person?
• add link to data archive once model is there

Reviewer Checklist

Formatting

• the pre-build checks succeed

General code quality

• naming conventions are met (see .clang-tidy for detailed information)
• no static analyzer warnings in new code parts (e.g., use clang-tidy for checking)

General usability

• old versions of pet-files are still loadable

Only if changes in UI

• new elements are also saved and loaded from pet-file
• check if tab order is still correct
• all new SpinBoxes are promoted
• new keybindings added to Petrack::keyBindings()