Module recognize_text 

Available on crate feature recognize_text only.

TextRecognizer — wraps VNRecognizeTextRequest for image-file OCR.

Structs

BoundingBox

Bounding box in normalised (0.0..=1.0) image coordinates with origin at the bottom-left (Vision convention — flip y if you want top-left origin).

RecognizedText

One recognised text observation.

RecognizedTextCandidate

Dedicated VNRecognizedText wrapper.

TextRecognizer

OCR engine.

Enums

RecognitionLevel

Recognition strategy passed to Vision. Fast is real-time-friendly, Accurate does layout analysis and is significantly slower.