Build OCR Features in C# Using IronOCR

Build OCR Features in C# Using IronOCR

Optical Character Recognition (OCR) enables applications to extract text from images and scanned documents. IronOCR is a popular C# library that simplifies OCR integration for .NET developers. This article shows how to set up IronOCR, implement common OCR features, and optimize results for accuracy and performance.

Prerequisites

• .NET 6 or later (reasonable default)
• Visual Studio or another C# IDE
• NuGet package manager
• Sample image or PDF to test

Installation

1. Create or open a .NET project.
2. Install IronOCR from NuGet:

bash
Copy CodeCopied
dotnet add package IronOcr 

3. Add using directive in your code files:

csharp
Copy CodeCopied
using IronOcr;

Basic OCR: Extract Text from an Image

1. Initialize the OCR engine and read an image:

csharp
Copy CodeCopied
var Ocr = new IronTesseract();
using (var Input = new OcrInput(“invoice.jpg”))
{
var Result = Ocr.Read(Input);
    Console.WriteLine(Result.Text);
}

2. Explanation:

• IronTesseract wraps Tesseract with pre-configured settings.
• OcrInput accepts file paths, byte arrays, URLs, and streams.

OCR from PDF

csharp
Copy CodeCopied
var Ocr = new IronTesseract();
using (var Input = new OcrInput(“scan.pdf”))
{
    Input.Deskew();           // optional preprocessing
    Input.EnhanceContrast();  // optional preprocessing
    var Result = Ocr.Read(Input);
    Console.WriteLine(Result.Text);
}

IronOCR handles multi-page PDFs and returns aggregated text.

Extract Structured Data (e.g., Invoice Fields)

1. Use regular expressions or simple parsing against Result.Text:

csharp
Copy CodeCopied
var text = Result.Text;
var invoiceNumber = Regex.Match(text, @“Invoice\s#?:?\s(\w+)”).Groups[1].Value;
var total = Regex.Match(text, @“Total\s[:\$]?\s([\d,.]+)”).Groups[1].Value;

2. For more robust extraction, combine OCR with template matching or ML-based parsers.

Improve Accuracy with Image Preprocessing

• Deskew(), EnhanceContrast(), RemoveNoise(), Sharpen() are available on OcrInput:

csharp
Copy CodeCopied
Input.Deskew()
     .EnhanceContrast()
     .RemoveNoise()
     .Sharpen();

• Choose preprocessing steps empirically per document type.

Language and Character Support

• Set specific OCR language packs if needed:

csharp
Copy CodeCopied
Ocr.Language = OcrLanguage.English;

• For other languages, load appropriate language data via IronOCR settings or NuGet language packs.

Handling Layout and Zones

• Use regions to focus OCR on parts of an image:

csharp
Copy CodeCopied
Input.AddImage(“form.jpg”);
Input.AddRectangle(100, 200, 400, 100); // x, y, width, height
var Result = Ocr.Read(Input);

• Process multiple zones separately to preserve structure.

Performance and Concurrency

• Reuse IronTesseract instance across calls to reduce model load time.
• For high throughput, run OCR tasks in parallel but limit degree of parallelism to avoid CPU/IO contention.

Output Formats

• Export OCR results to plain text, searchable PDF, or JSON:

csharp
Copy CodeCopied
File.WriteAllText(“output.txt”, Result.Text);
Result.SaveAsSearchablePdf(“searchable.pdf”);

Error Handling and Logging

• Catch exceptions and inspect Result.PageErrors or Result.Paragraphs for diagnostics.

csharp
Copy CodeCopied
try { var Result = Ocr.Read(Input); }
catch (Exception ex) { Console.WriteLine(ex.Message); }

Example: Console App Putting It Together

csharp
Copy CodeCopied
using System;
using IronOcr;
using System.Text.RegularExpressions;

class Program
{
    static void Main()
    {
        var Ocr = new IronTesseract();
        using (var Input = new OcrInput(“invoice.jpg”))
        {
            Input.Deskew().EnhanceContrast().RemoveNoise();
            var Result = Ocr.Read(Input);

            Console.WriteLine(“Raw Text:\n” + Result.Text);

            var invoiceNumber = Regex.Match(Result.Text, @“Invoice\s#?:?\s(\w+)”).Groups[1].Value;
            var total = Regex.Match(Result.Text, @“Total\s[:\$]?\s([\d,.]+)”).Groups[1].Value;

            Console.WriteLine(\("Invoice #: </span><span class="token interpolation-string interpolation" style="color: rgb(57, 58, 52);">{</span><span class="token interpolation-string interpolation expression language-csharp">invoiceNumber</span><span class="token interpolation-string interpolation" style="color: rgb(57, 58, 52);">}</span><span class="token interpolation-string" style="color: rgb(163, 21, 21);">"</span><span class="token" style="color: rgb(57, 58, 52);">)</span><span class="token" style="color: rgb(57, 58, 52);">;</span><span> </span><span>            Console</span><span class="token" style="color: rgb(57, 58, 52);">.</span><span class="token" style="color: rgb(57, 58, 52);">WriteLine</span><span class="token" style="color: rgb(57, 58, 52);">(</span><span class="token interpolation-string" style="color: rgb(163, 21, 21);">\)“Total: {total}”);

            Result.SaveAsSearchablePdf(“invoice_searchable.pdf”);
        }
    }
}

Tips and Best Practices

• Use good-quality scans (300 DPI recommended).
• Preprocess images to remove skew, noise, and improve contrast.
• Restrict OCR to zones when possible to reduce false positives.
• Reuse engine instances and cache results for frequent documents.
• Validate parsed fields with regex or rules to catch OCR errors.

Further Reading

• IronOCR official docs and API reference for advanced options (e.g., custom training, image filters).
• Tesseract OCR concepts for understanding recognition limits and language training.

This guide provides a practical starting point to add OCR features to C# apps using IronOCR. Adjust preprocessing and parsing strategies to your document types for best results.