Multithreaded Tesseract OCR

using IronOcr;

var ocrTesseract = new IronTesseract();

// Optional: cap how many pages are OCR'd at once to bound peak memory.
// Defaults to Environment.ProcessorCount. Set MultiThreaded = false to read sequentially.
ocrTesseract.MaxDegreeOfParallelism = 4;

using var ocrInput = new OcrInput();
ocrInput.LoadPdf("scan.pdf");

// Image processing is automatically multi-threaded
ocrInput.Deskew();

// OCR reading is automatically multi-threaded too
var ocrResult = ocrTesseract.Read(ocrInput);

PM > Install-Package IronOcr

Install-Package IronOcr

Multithreaded Tesseract OCR

All IronOCR image processing and OCR reading operations are multithreaded by default, so no special API is needed to get started. The legacy ReadMultithreaded method is no longer required.

IronTesseract will automatically attempt to use all threads available on all cores and will consider responsiveness on the main / GUI thread elegantly.

When you need to bound that parallelism, two properties on IronTesseract control it. MaxDegreeOfParallelism caps how many pages are OCR'd concurrently within a single Read call - each concurrent page runs its own native Tesseract engine, so lowering the value lowers peak memory. Setting MultiThreaded to false reads pages sequentially for the lowest and most predictable memory footprint, and takes precedence over MaxDegreeOfParallelism.

How to Use Tesseract Multi-threading

Install an OCR library to use Tesseract multi-threading
Make an IronTesseract object.
Use LoadImage to construct an OcrInput and provide the image path into an object
Invoke all necessary methods for image processing.
Use the Read method to read the text

Learn Async OCR Techniques with IronOCR!

View on GitHub