PdfNetStack.Ocr namespace
Package: PdfNetStack.Ocr
Namespaces
Section titled “Namespaces”Classes
Section titled “Classes”An 8 bit gray image: rows top to bottom, Width bytes each, 0 black.
Text found in one image, lines in reading order.
A line of text.
What to read.
A word.
OCR with the PP-OCRv5 mobile models (PaddleOCR, Apache-2.0) on ONNX Runtime: a DB text detector, a line orientation classifier and CTC recognizers for Latin and Arabic script. Pre and post processing are managed code; the native runtime only sees float tensors.
Settings for OnnxOcrEngine.
Makes scanned PDFs searchable: finds the image that is each page’s scan, reads it with an IOcrEngine, and lays the words over it as invisible text, so the page looks the same but selects, copies and searches.
Settings for PdfOcr.
The result for one page. Coordinates are in PDF user space: points, y up, before /Rotate.
The result for a document.
Interfaces
Section titled “Interfaces”Recognizes text in images. Implementations must be safe to call from several threads.
Renders a page to pixels for OCR, for pages PdfOcr cannot read from their images. PdfNetStack does not render; an application supplies a renderer, for example one built on PDFium, through Rasterizer.
What happened to a page.