The OCR interface is responsible for recognizing text and layout from a rasterized page.
Purpose
OCR is used when the document lacks a real text layer or when the page needs to be interpreted from a rendered image.
The OCR seam: recognise a page into words, lines and positions with per-word confidence, which is what makes provenance and scoring possible.