pre-process pages or images with an orientation-detection step

[atelierdesdonnees]

Hello,
Is there any plan to preprocess pages or zones passed to the OCR engine with an orientation-detection step, like tesseract's image_to_osd function ?
Unless it would be more interesting to use detectron2 to perform orientation detection...