Text detection

Is your feature request related to a problem? Please describe.

I'm looking for a mechanism to resize images that contain text overlays.

Describe the solution you'd like

I would love if the mechanism for preserving faces could apply to OCR detected text.

Additional context