xImage.OCR vs IronOCR: .NET OCR Library Comparison

Language Package Fragmentation in Practice

The consequence of one-package-per-language manifests concretely in your project file. For an application that processes documents in five European languages, the package manifest looks like this:

<PackageReference Include="RasterEdge.XImage.OCR" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.English" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.German" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.French" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.Spanish" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.Italian" Version="12.4.0" />

<PackageReference Include="RasterEdge.XImage.OCR" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.English" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.German" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.French" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.Spanish" Version="12.4.0" />
<PackageReference Include="XImage.OCR.Language.Italian" Version="12.4.0" />

Version mismatches between core and language packages cause runtime initialization errors. When RasterEdge.XImage.OCR is at 12.4.0 and XImage.OCR.Language.English is at 12.3.0 — a realistic scenario after any partial update — the library fails at runtime with errors that do not clearly identify version sync as the cause. Your CI/CD pipeline now has 6 distinct restore operations and 6 independent failure points. Adding Asian languages for a global deployment adds CJK packages: Chinese Simplified, Chinese Traditional, Japanese, and Korean each require their own package, bringing a 10-language application to 11 separate NuGet dependencies.

Detailed Feature Comparison

XImage.OCR Approach

XImage.OCR distributes each language as a separate NuGet package. The core installation command installs the OCR engine only. You then install each required language:

# XImage.OCR: Install sequence for a multilingual application
dotnet add package RasterEdge.XImage.OCR
dotnet add package XImage.OCR.Language.English
dotnet add package XImage.OCR.Language.German
dotnet add package XImage.OCR.Language.French
dotnet add package XImage.OCR.Language.Spanish
dotnet add package XImage.OCR.Language.Italian
dotnet add package XImage.OCR.Language.Portuguese
dotnet add package XImage.OCR.Language.ChineseSimplified
dotnet add package XImage.OCR.Language.ChineseTraditional
dotnet add package XImage.OCR.Language.Japanese
# 10 languages = 11 package commands

# XImage.OCR: Install sequence for a multilingual application
dotnet add package RasterEdge.XImage.OCR
dotnet add package XImage.OCR.Language.English
dotnet add package XImage.OCR.Language.German
dotnet add package XImage.OCR.Language.French
dotnet add package XImage.OCR.Language.Spanish
dotnet add package XImage.OCR.Language.Italian
dotnet add package XImage.OCR.Language.Portuguese
dotnet add package XImage.OCR.Language.ChineseSimplified
dotnet add package XImage.OCR.Language.ChineseTraditional
dotnet add package XImage.OCR.Language.Japanese
# 10 languages = 11 package commands

The code itself reflects the same complexity. Setting up multi-language OCR requires that all language packages are already installed — the library cannot validate this at compile time:

// XImage.OCR: Language codes must match installed packages exactly
// If XImage.OCR.Language.German is not installed, this fails at runtime

var ocrHandler = new OCRHandler();

// License activation required before any OCR operation
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-license-key");

// Set multiple languages — each requires its own NuGet package installed
ocrHandler.Languages = new[] { "eng", "deu", "fra", "spa", "ita" };

string result = ocrHandler.Process(imagePath);

// XImage.OCR: Language codes must match installed packages exactly
// If XImage.OCR.Language.German is not installed, this fails at runtime

var ocrHandler = new OCRHandler();

// License activation required before any OCR operation
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-license-key");

// Set multiple languages — each requires its own NuGet package installed
ocrHandler.Languages = new[] { "eng", "deu", "fra", "spa", "ita" };

string result = ocrHandler.Process(imagePath);

' XImage.OCR: Language codes must match installed packages exactly
' If XImage.OCR.Language.German is not installed, this fails at runtime

Dim ocrHandler As New OCRHandler()

' License activation required before any OCR operation
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-license-key")

' Set multiple languages — each requires its own NuGet package installed
ocrHandler.Languages = New String() {"eng", "deu", "fra", "spa", "ita"}

Dim result As String = ocrHandler.Process(imagePath)

If a language package is missing or on the wrong version, the failure happens at runtime during the OCR call, not at compile time or during package restore. In a deployment pipeline, this means discovering the problem in production or pre-production environments rather than on the developer's machine.

IronOCR Approach

IronOCR installs as a single package. Language support for common Western languages is bundled by default; additional language packs install as optional packages following the same one-package pattern, but you start from a working state rather than a broken one:

# IronOCR: Install everything in one command
dotnet add package IronOcr
# English included. Add specific language packs only when needed.
dotnet add package IronOcr.Languages.German

# IronOCR: Install everything in one command
dotnet add package IronOcr
# English included. Add specific language packs only when needed.
dotnet add package IronOcr.Languages.German

The code for multi-language OCR uses type-safe OcrLanguage enum values, not string codes, which eliminates the class of runtime errors that come from mistyped language identifiers:

// IronOCR: Type-safe languages, single package
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

var ocr = new IronTesseract();
ocr.Language = OcrLanguage.English;
ocr.AddSecondaryLanguage(OcrLanguage.German);
ocr.AddSecondaryLanguage(OcrLanguage.French);

var result = ocr.Read("multilingual-document.jpg");
Console.WriteLine(result.Text);

// IronOCR: Type-safe languages, single package
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

var ocr = new IronTesseract();
ocr.Language = OcrLanguage.English;
ocr.AddSecondaryLanguage(OcrLanguage.German);
ocr.AddSecondaryLanguage(OcrLanguage.French);

var result = ocr.Read("multilingual-document.jpg");
Console.WriteLine(result.Text);

Imports IronOcr

' IronOCR: Type-safe languages, single package
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"

Dim ocr As New IronTesseract()
ocr.Language = OcrLanguage.English
ocr.AddSecondaryLanguage(OcrLanguage.German)
ocr.AddSecondaryLanguage(OcrLanguage.French)

Dim result = ocr.Read("multilingual-document.jpg")
Console.WriteLine(result.Text)

The IronTesseract setup guide covers language configuration in detail. The multiple languages how-to walks through combining primary and secondary languages for documents with mixed-language content.

For a 10-language application: XImage.OCR requires 11 packages, 11 lines in .csproj, and ongoing version synchronization across all 11 when updates arrive. IronOCR requires 1 package in .csproj, plus optional per-language packs when needed. That difference compounds across teams.

XImage.OCR Approach

XImage.OCR passes images directly to Tesseract. A scanned invoice with 2-degree skew, scanner noise, and 150 DPI resolution gets processed as-is. The OCRHandler has no methods for image correction:

// XImage.OCR: No preprocessing available
// A skewed, noisy, low-DPI scan goes straight to Tesseract

var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";

// Direct OCR on problematic image — skew, noise, and low resolution
// all degrade accuracy with no mitigation available in the library
string result = ocrHandler.Process(imagePath);

// To improve results, you would need:
// - External library (OpenCV.NET, ImageSharp, etc.)
// - 100-200 lines of deskew and noise-reduction code
// - Per-document-type tuning
// - Image processing expertise your OCR developer may not have

// XImage.OCR: No preprocessing available
// A skewed, noisy, low-DPI scan goes straight to Tesseract

var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";

// Direct OCR on problematic image — skew, noise, and low resolution
// all degrade accuracy with no mitigation available in the library
string result = ocrHandler.Process(imagePath);

// To improve results, you would need:
// - External library (OpenCV.NET, ImageSharp, etc.)
// - 100-200 lines of deskew and noise-reduction code
// - Per-document-type tuning
// - Image processing expertise your OCR developer may not have

' XImage.OCR: No preprocessing available
' A skewed, noisy, low-DPI scan goes straight to Tesseract

Dim ocrHandler As New OCRHandler()
ocrHandler.Language = "eng"

' Direct OCR on problematic image — skew, noise, and low resolution
' all degrade accuracy with no mitigation available in the library
Dim result As String = ocrHandler.Process(imagePath)

' To improve results, you would need:
' - External library (OpenCV.NET, ImageSharp, etc.)
' - 100-200 lines of deskew and noise-reduction code
' - Per-document-type tuning
' - Image processing expertise your OCR developer may not have

The workaround is pulling in a separate image processing library, writing preprocessing code yourself, and maintaining that code across library updates. For teams already deep in RasterEdge's ecosystem, some tooling is available through other RasterEdge SDK components — but that requires purchasing and integrating additional products.

IronOCR Approach

IronOCR applies preprocessing as chainable methods on OcrInput, between document loading and OCR execution. No external library is needed:

// IronOCR: Built-in preprocessing pipeline
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

using var input = new OcrInput();
input.LoadImage("low-quality-scan.jpg");

// Apply corrections in sequence
input.Deskew();              // Detect and correct rotation angle
input.DeNoise();             // Remove scanner artifacts and noise
input.Contrast();            // Enhance contrast for faded text
input.Binarize();            // Convert to clean black-and-white
input.EnhanceResolution(300); // Scale low-DPI images to 300 DPI

var result = new IronTesseract().Read(input);
Console.WriteLine(result.Text);
Console.WriteLine($"Confidence: {result.Confidence}%");

// IronOCR: Built-in preprocessing pipeline
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

using var input = new OcrInput();
input.LoadImage("low-quality-scan.jpg");

// Apply corrections in sequence
input.Deskew();              // Detect and correct rotation angle
input.DeNoise();             // Remove scanner artifacts and noise
input.Contrast();            // Enhance contrast for faded text
input.Binarize();            // Convert to clean black-and-white
input.EnhanceResolution(300); // Scale low-DPI images to 300 DPI

var result = new IronTesseract().Read(input);
Console.WriteLine(result.Text);
Console.WriteLine($"Confidence: {result.Confidence}%");

Imports IronOcr

' IronOCR: Built-in preprocessing pipeline
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"

Using input As New OcrInput()
    input.LoadImage("low-quality-scan.jpg")

    ' Apply corrections in sequence
    input.Deskew()              ' Detect and correct rotation angle
    input.DeNoise()             ' Remove scanner artifacts and noise
    input.Contrast()            ' Enhance contrast for faded text
    input.Binarize()            ' Convert to clean black-and-white
    input.EnhanceResolution(300) ' Scale low-DPI images to 300 DPI

    Dim result = New IronTesseract().Read(input)
    Console.WriteLine(result.Text)
    Console.WriteLine($"Confidence: {result.Confidence}%")
End Using

The image quality correction guide covers the full preprocessing filter set. For orientation-specific corrections, the image orientation correction guide and page rotation detection handle multi-angle documents.

The practical impact of preprocessing on real-world documents is significant. Tesseract without preprocessing on a scanned document at 150 DPI with 5-degree skew produces accuracy in the 60-75% range. With deskew and resolution enhancement, the same document at 300 DPI with corrected rotation reaches 93-97% accuracy. XImage.OCR cannot close that gap without external code. IronOCR closes it with 5 method calls.

The Free Alternative Gap

The charlesw/tesseract wrapper on NuGet has 8 million+ downloads, active maintenance, and Apache 2.0 licensing — free for commercial use. It provides essentially the same OCR output as XImage.OCR because both are calling the same Tesseract engine. The differences are:

• Language support: Free tessdata has 100+ languages; XImage.OCR supports ~15 as paid packages
• Community: charlesw/tesseract has 3,000+ Stack Overflow questions and extensive community documentation; XImage.OCR has fewer than 100
• Preprocessing: Neither library provides built-in preprocessing — both pass images directly to Tesseract
• Threading: Both have the same thread-safety constraint (per-thread instances required)

When the commercial product has fewer languages, smaller community, the same preprocessing gap, and the same threading model as the free alternative, the commercial value proposition becomes hard to articulate beyond vendor support and RasterEdge ecosystem integration.

IronOCR as a Commercial Alternative

IronOCR justifies its commercial pricing through capabilities that neither XImage.OCR nor free wrappers provide: automated preprocessing, native PDF support, thread-safe design, and cross-platform deployment. The $999 Lite license is a one-time perpetual purchase. The question is not whether commercial OCR is worth paying for — it often is. The question is what you get for the money.

// What the IronOCR Lite license buys vs what commercial pricing buys with XImage.OCR

// IronOCR: Preprocessing, PDF, thread safety, cross-platform in one call
var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadPdf("scanned-invoice.pdf");   // Native PDF — no extra SDK
input.Deskew();                          // Built-in preprocessing
input.DeNoise();                         // No external library needed
var result = ocr.Read(input);
result.SaveAsSearchablePdf("output.pdf"); // Searchable PDF output
Console.WriteLine($"Confidence: {result.Confidence}%");

// XImage.OCR: Same Tesseract core, no preprocessing, no PDF, not thread-safe
// Requires purchasing and integrating RasterEdge PDF SDK for PDF operations

// What the IronOCR Lite license buys vs what commercial pricing buys with XImage.OCR

// IronOCR: Preprocessing, PDF, thread safety, cross-platform in one call
var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadPdf("scanned-invoice.pdf");   // Native PDF — no extra SDK
input.Deskew();                          // Built-in preprocessing
input.DeNoise();                         // No external library needed
var result = ocr.Read(input);
result.SaveAsSearchablePdf("output.pdf"); // Searchable PDF output
Console.WriteLine($"Confidence: {result.Confidence}%");

// XImage.OCR: Same Tesseract core, no preprocessing, no PDF, not thread-safe
// Requires purchasing and integrating RasterEdge PDF SDK for PDF operations

Imports IronOcr

' What the IronOCR Lite license buys vs what commercial pricing buys with XImage.OCR

' IronOCR: Preprocessing, PDF, thread safety, cross-platform in one call
Dim ocr As New IronTesseract()
Using input As New OcrInput()
    input.LoadPdf("scanned-invoice.pdf")   ' Native PDF — no extra SDK
    input.Deskew()                         ' Built-in preprocessing
    input.DeNoise()                        ' No external library needed
    Dim result = ocr.Read(input)
    result.SaveAsSearchablePdf("output.pdf") ' Searchable PDF output
    Console.WriteLine($"Confidence: {result.Confidence}%")
End Using

' XImage.OCR: Same Tesseract core, no preprocessing, no PDF, not thread-safe
' Requires purchasing and integrating RasterEdge PDF SDK for PDF operations

The reading text from images tutorial and C# Tesseract OCR guide document the full capability set. For teams evaluating why IronOCR is preferable to a basic Tesseract wrapper specifically, the dedicated Why IronOCR and not Tesseract page covers the technical differentiators in detail.

XImage.OCR Approach

XImage.OCR does not natively process PDFs. The workflow requires the RasterEdge PDF SDK — a separate commercial purchase — to render PDF pages to images, which XImage.OCR then processes one page at a time via temporary files:

// XImage.OCR: PDF requires RasterEdge PDF SDK (additional purchase)
// Workflow: PDF -> render to image -> temp file -> OCR -> delete temp file

var pdfDocument = new PDFDocument(pdfPath); // Requires separate RasterEdge SDK

var results = new System.Text.StringBuilder();
var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";

for (int i = 0; i < pdfDocument.PageCount; i++)
{
    // Render PDF page to image at 200 DPI
    var pageImage = pdfDocument.RenderPage(i, 200);

    string tempPath = Path.GetTempFileName() + ".png";
    pageImage.Save(tempPath);

    try
    {
        string pageText = ocrHandler.Process(tempPath);
        results.AppendLine($"--- Page {i + 1} ---");
        results.AppendLine(pageText);
    }
    finally
    {
        File.Delete(tempPath); // Manual cleanup required
    }
}

return results.ToString();

// XImage.OCR: PDF requires RasterEdge PDF SDK (additional purchase)
// Workflow: PDF -> render to image -> temp file -> OCR -> delete temp file

var pdfDocument = new PDFDocument(pdfPath); // Requires separate RasterEdge SDK

var results = new System.Text.StringBuilder();
var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";

for (int i = 0; i < pdfDocument.PageCount; i++)
{
    // Render PDF page to image at 200 DPI
    var pageImage = pdfDocument.RenderPage(i, 200);

    string tempPath = Path.GetTempFileName() + ".png";
    pageImage.Save(tempPath);

    try
    {
        string pageText = ocrHandler.Process(tempPath);
        results.AppendLine($"--- Page {i + 1} ---");
        results.AppendLine(pageText);
    }
    finally
    {
        File.Delete(tempPath); // Manual cleanup required
    }
}

return results.ToString();

Imports System.Text
Imports System.IO

' XImage.OCR: PDF requires RasterEdge PDF SDK (additional purchase)
' Workflow: PDF -> render to image -> temp file -> OCR -> delete temp file

Dim pdfDocument As New PDFDocument(pdfPath) ' Requires separate RasterEdge SDK

Dim results As New StringBuilder()
Dim ocrHandler As New OCRHandler()
ocrHandler.Language = "eng"

For i As Integer = 0 To pdfDocument.PageCount - 1
    ' Render PDF page to image at 200 DPI
    Dim pageImage = pdfDocument.RenderPage(i, 200)

    Dim tempPath As String = Path.GetTempFileName() & ".png"
    pageImage.Save(tempPath)

    Try
        Dim pageText As String = ocrHandler.Process(tempPath)
        results.AppendLine($"--- Page {i + 1} ---")
        results.AppendLine(pageText)
    Finally
        File.Delete(tempPath) ' Manual cleanup required
    End Try
Next

Return results.ToString()

This pattern introduces temp file management, explicit cleanup, and dependency on a second commercial product. The 200 DPI render target in this pattern also falls below the 300 DPI threshold where Tesseract accuracy is optimal, which further impacts quality. Teams using XImage.OCR for PDF workflows are effectively paying for two products to do what one covers elsewhere.

IronOCR Approach

IronOCR handles PDFs natively. No intermediate steps, no temp files, no second SDK:

// IronOCR: Native PDF input, no extra dependencies
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

using var input = new OcrInput();
input.LoadPdf("scanned-invoice.pdf");  // Direct PDF loading
var result = new IronTesseract().Read(input);
result.SaveAsSearchablePdf("searchable-output.pdf");
Console.WriteLine(result.Text);

// IronOCR: Native PDF input, no extra dependencies
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

using var input = new OcrInput();
input.LoadPdf("scanned-invoice.pdf");  // Direct PDF loading
var result = new IronTesseract().Read(input);
result.SaveAsSearchablePdf("searchable-output.pdf");
Console.WriteLine(result.Text);

Imports IronOcr

' IronOCR: Native PDF input, no extra dependencies
IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"

Using input As New OcrInput()
    input.LoadPdf("scanned-invoice.pdf")  ' Direct PDF loading
    Dim result = New IronTesseract().Read(input)
    result.SaveAsSearchablePdf("searchable-output.pdf")
    Console.WriteLine(result.Text)
End Using

Password-protected PDFs add one parameter:

using var input = new OcrInput();
input.LoadPdf("encrypted.pdf", Password: "secret");
var result = new IronTesseract().Read(input);

using var input = new OcrInput();
input.LoadPdf("encrypted.pdf", Password: "secret");
var result = new IronTesseract().Read(input);

Imports IronOcr

Using input As New OcrInput()
    input.LoadPdf("encrypted.pdf", Password: "secret")
    Dim result = New IronTesseract().Read(input)
End Using

The PDF input guide covers multi-page PDFs, page range selection, and DPI configuration. The searchable PDF guide documents output options. For teams building document archival pipelines, the PDF OCR in C# use case page covers the complete workflow.

Package Management Has Become a Maintenance Burden

Teams that started with two or three XImage.OCR language packages often reach a tipping point when the application expands to 8 or 10 languages. The project file now has 9–11 package references all requiring version synchronization. The developer who updated only the core package silently broke OCR for every language. The CI/CD pipeline intermittently fails because the package cache restored an old language pack. At this stage, the version-sync overhead exceeds any value the per-language packaging provides, and consolidating to a single-package solution becomes the obvious architectural decision.

Real-World Document Quality Demands Preprocessing

Teams deploying XImage.OCR against controlled, high-quality scans initially encounter no accuracy problems. The issue surfaces when the document pipeline expands — mobile photo uploads, legacy archive scans, fax documents, forms filled in by hand and scanned at 150 DPI. Without preprocessing, Tesseract accuracy on these inputs is poor regardless of which wrapper you use. Teams then face a choice: build and maintain their own preprocessing library integration, or move to a solution with preprocessing built in. IronOCR's built-in deskew, denoise, and resolution enhancement eliminate the external dependency entirely and deliver the accuracy improvement without requiring image processing expertise on the OCR developer's part.

PDF Workflows Require a Second SDK Purchase

When a team's document pipeline expands from image files to PDFs — the most common escalation path for document processing applications — XImage.OCR forces a second RasterEdge product purchase. The PDF SDK is a separate commercial license. Teams paying commercial pricing for OCR did not budget for a second commercial license to make the OCR work on the most common document format in enterprise environments. IronOCR's native PDF support makes this concern irrelevant from day one, and the PDF OCR example demonstrates exactly how direct the integration is.

Cross-Platform Deployment Becomes a Requirement

XImage.OCR's Windows-primary architecture becomes a hard constraint when containerized deployment on Linux or a macOS development environment enters the picture. Modern .NET development teams run Docker for local development and deploy to Linux-based container infrastructure. A Windows-only OCR library cannot participate in that stack. IronOCR supports Windows, Linux, macOS, and Docker from the same package with no configuration changes required — the Docker deployment guide and Linux deployment guide cover the setup in detail.

Thread Safety Limits Throughput in Server Applications

OCR in a server application, processing batch jobs, or handling concurrent HTTP requests requires multi-threading. XImage.OCR's per-thread handler model means each concurrent thread loads its own Tesseract engine instance, consuming 40–100MB per language per thread. A 4-thread server processing documents in 5 languages loads approximately 900MB–1.2GB just for OCR handler instances. IronOCR's thread-safe design shares a single instance across all threads — that single instance handles concurrent requests without multiplying memory consumption. Teams hitting memory or throughput walls in production often trace the root cause to this architectural difference.

Replacing OCRHandler with IronTesseract

The primary code substitution is straightforward. Replace OCRHandler with IronTesseract, replace string language codes ("eng", "deu") with type-safe OcrLanguage enum values, and wrap image loading in OcrInput. License activation moves from LicenseManager.SetLicense() to the IronOcr.License.LicenseKey static property, which can be set from an environment variable for deployment safety:

// Before: XImage.OCR
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-key");
var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";
string text = ocrHandler.Process(imagePath);

// After: IronOCR
IronOcr.License.LicenseKey = Environment.GetEnvironmentVariable("IRONOCR_LICENSE");
var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadImage(imagePath);
var result = ocr.Read(input);
string text = result.Text;

// Before: XImage.OCR
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-key");
var ocrHandler = new OCRHandler();
ocrHandler.Language = "eng";
string text = ocrHandler.Process(imagePath);

// After: IronOCR
IronOcr.License.LicenseKey = Environment.GetEnvironmentVariable("IRONOCR_LICENSE");
var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadImage(imagePath);
var result = ocr.Read(input);
string text = result.Text;

Imports System
Imports IronOcr

' Before: XImage.OCR
RasterEdge.XImage.OCR.License.LicenseManager.SetLicense("your-key")
Dim ocrHandler As New OCRHandler()
ocrHandler.Language = "eng"
Dim text As String = ocrHandler.Process(imagePath)

' After: IronOCR
IronOcr.License.LicenseKey = Environment.GetEnvironmentVariable("IRONOCR_LICENSE")
Dim ocr As New IronTesseract()
Using input As New OcrInput()
    input.LoadImage(imagePath)
    Dim result = ocr.Read(input)
    text = result.Text
End Using

The full image input guide covers all input types: file paths, byte arrays, streams, and URLs.

Adding Preprocessing Where None Existed

Teams migrating from XImage.OCR have an immediate opportunity to add preprocessing that was unavailable before. The migration is not just a class-name swap — it is a chance to fix accuracy problems that were silently accepted because no better option existed. After loading input, add the filters appropriate for your document types:

using var input = new OcrInput();
input.LoadImage(imagePath);

// Add these lines immediately after loading — no other changes needed
input.Deskew();
input.DeNoise();
input.EnhanceResolution(300);

var result = new IronTesseract().Read(input);

using var input = new OcrInput();
input.LoadImage(imagePath);

// Add these lines immediately after loading — no other changes needed
input.Deskew();
input.DeNoise();
input.EnhanceResolution(300);

var result = new IronTesseract().Read(input);

Imports IronOcr

Using input As New OcrInput()
    input.LoadImage(imagePath)

    ' Add these lines immediately after loading — no other changes needed
    input.Deskew()
    input.DeNoise()
    input.EnhanceResolution(300)

    Dim result = New IronTesseract().Read(input)
End Using

The image filters tutorial provides guidance on which filters to apply for specific document quality issues. Apply them only to documents where quality is variable — clean, high-resolution scans do not benefit from preprocessing and the additional step adds processing time.

Simplifying the Package List

Remove all XImage.OCR packages from the project. The package removal is the longest part of the process for applications with many languages installed:

# Remove all XImage.OCR packages
dotnet remove package RasterEdge.XImage.OCR
dotnet remove package XImage.OCR.Language.English
dotnet remove package XImage.OCR.Language.German
dotnet remove package XImage.OCR.Language.French
# ... remove all language packages

# Add IronOCR
dotnet add package IronOcr

# Remove all XImage.OCR packages
dotnet remove package RasterEdge.XImage.OCR
dotnet remove package XImage.OCR.Language.English
dotnet remove package XImage.OCR.Language.German
dotnet remove package XImage.OCR.Language.French
# ... remove all language packages

# Add IronOCR
dotnet add package IronOcr

Update the CI/CD pipeline to remove the multiple restore operations and simplify the package dependency verification steps. The version synchronization logic — if it was automated — can be deleted entirely.

Handling Region-Based OCR

XImage.OCR exposes region processing via ProcessRegion(imagePath, rectangle). IronOCR handles this through CropRectangle passed at input loading time:

// IronOCR region-based OCR
var region = new CropRectangle(x: 0, y: 0, width: 600, height: 100);
using var input = new OcrInput();
input.LoadImage("invoice.jpg", region);
var text = new IronTesseract().Read(input).Text;

// IronOCR region-based OCR
var region = new CropRectangle(x: 0, y: 0, width: 600, height: 100);
using var input = new OcrInput();
input.LoadImage("invoice.jpg", region);
var text = new IronTesseract().Read(input).Text;

' IronOCR region-based OCR
Dim region As New CropRectangle(x:=0, y:=0, width:=600, height:=100)
Using input As New OcrInput()
    input.LoadImage("invoice.jpg", region)
    Dim text As String = New IronTesseract().Read(input).Text
End Using

The region-based OCR guide and region crop example document the approach for fixed-layout documents like invoices and forms.