TPJ.OCR 10.0.0

TPJ.OCR

Simple OCR helper library for .NET 10 built on top of Tesseract.

What it does

TPJ.OCR makes it easy to:

• read text from a full image
• read text from a specific rectangle in an image
• read multiple named areas in one call
• read from a file path, Stream, or byte[]

Requirements

• .NET 10
• Tesseract language data in a tessdata folder

By default, the library looks for tessdata in the application output folder:

• tessdata/eng.traineddata

If your files are stored somewhere else, set OCROptions.DataPath.

Install

NuGet

dotnet add package TPJ.OCR

Project reference

<ItemGroup>
  <ProjectReference Include="..\TPJ.OCR\TPJ.OCR.csproj" />
</ItemGroup>

Namespaces

using TPJ.OCR;
using TPJ.OCR.Enums;

Simple examples

Read all text from an image

var text = OCRReader.ReadText("ExampleImage.png");
Console.WriteLine(text);

Read digits from a specific area

Use ContentType.Digits when the area should only contain numbers.

var employeeNumber = OCRReader.ReadText(
    "ExampleImage.png",
    ContentType.Digits,
    new OCRRectangle(460, 220, 240, 100));

Console.WriteLine(employeeNumber);

Read multiple named areas

var results = OCRReader.ReadAreas(
    "ExampleImage.png",
    new List<OCRArea>
    {
        OCRArea.FromBounds("Name", 300, 40, 400, 100, ContentType.Alphabetic),
        OCRArea.FromBounds("Payroll Name", 1350, 40, 350, 100, ContentType.Alphabetic),
        OCRArea.FromBounds("HRMS No", 460, 220, 240, 100, ContentType.Digits),
        OCRArea.FromBounds("Pay Date", 1170, 220, 540, 100)
    });

foreach (var item in results)
{
    Console.WriteLine($"{item.AreaName} = {item.Value}");
}

Read multiple areas into a dictionary

var values = OCRReader.ReadAreaMap(
    "ExampleImage.png",
    new[]
    {
        OCRArea.FromBounds("Name", 300, 40, 400, 100, ContentType.Alphabetic),
        OCRArea.FromBounds("HRMS No", 460, 220, 240, 100, ContentType.Digits)
    });

Console.WriteLine(values["Name"]);
Console.WriteLine(values["HRMS No"]);

Read from a stream

using var stream = File.OpenRead("ExampleImage.png");
var text = OCRReader.ReadText(stream);

Console.WriteLine(text);

Read from bytes

var imageBytes = File.ReadAllBytes("ExampleImage.png");
var text = OCRReader.ReadText(imageBytes);

Console.WriteLine(text);

Use custom OCR options

var options = new OCROptions
{
    DataPath = Path.Combine(AppContext.BaseDirectory, "tessdata"),
    Language = "eng",
    NormalizeWhitespace = true
};

var text = OCRReader.ReadText("ExampleImage.png", options);
Console.WriteLine(text);

Notes

• OCRRectangle(x, y, width, height) uses pixel coordinates.
• ContentType.Digits and ContentType.Alphabetic apply simple result cleanup for common OCR mistakes.
• OCR areas must fit inside the image bounds.

Example app

See TPJ.TestOCR/Program.cs for a working sample that reads:

• the full image
• a single area
• multiple areas