Google OCR & IronOCR 之間的比較

// charlesw/Tesseract: engine creation requires tessdata path at runtime
// The path must resolve correctly in every deployment target

private const string TessDataPath = @"./tessdata"; // Works on dev — breaks in Docker

public string ExtractText(string imagePath)
{
    // TesseractEngine P/Invokes into platform-specific native DLLs
    // Fails with DllNotFoundException if binaries are not in the right location
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img = Pix.LoadFromFile(imagePath);
    using var page = engine.Process(img);

    return page.GetText();
}

Imports Tesseract

Private Const TessDataPath As String = "./tessdata" ' Works on dev — breaks in Docker

Public Function ExtractText(imagePath As String) As String
    ' TesseractEngine P/Invokes into platform-specific native DLLs
    ' Fails with DllNotFoundException if binaries are not in the right location
    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(imagePath)
            Using page As Page = engine.Process(img)
                Return page.GetText()
            End Using
        End Using
    End Using
End Function

// Production code must handle path differences across environments
// There is no standard solution — every team builds their own

public class TesseractEngineFactory
{
    private static string ResolveTessDataPath()
    {
        // Dev machine: ./tessdata relative to executable
        // Docker: /app/tessdata (must be COPY'd into image)
        // Azure App Service: D:\home\site\wwwroot\tessdata
        // All three paths are different; all three need correct traineddata files

        var candidates = new[]
        {
            Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "tessdata"),
            Path.Combine(Directory.GetCurrentDirectory(), "tessdata"),
            "/app/tessdata",  // Docker-specific hard-code
        };

        foreach (var candidate in candidates)
        {
            if (Directory.Exists(candidate))
                return candidate;
        }

        throw new DirectoryNotFoundException("tessdata not found — deployment misconfigured");
    }

    public static TesseractEngine Create()
    {
        // TesseractEngine P/Invokes tesseract50.dll on Windows, libtesseract.so on Linux
        // If the native binary for the current runtime identifier is missing, throws DllNotFoundException
        return new TesseractEngine(ResolveTessDataPath(), "eng", EngineMode.Default);
    }
}

Imports System
Imports System.IO

' Production code must handle path differences across environments
' There is no standard solution — every team builds their own

Public Class TesseractEngineFactory
    Private Shared Function ResolveTessDataPath() As String
        ' Dev machine: ./tessdata relative to executable
        ' Docker: /app/tessdata (must be COPY'd into image)
        ' Azure App Service: D:\home\site\wwwroot\tessdata
        ' All three paths are different; all three need correct traineddata files

        Dim candidates = New String() {
            Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "tessdata"),
            Path.Combine(Directory.GetCurrentDirectory(), "tessdata"),
            "/app/tessdata"  ' Docker-specific hard-code
        }

        For Each candidate In candidates
            If Directory.Exists(candidate) Then
                Return candidate
            End If
        Next

        Throw New DirectoryNotFoundException("tessdata not found — deployment misconfigured")
    End Function

    Public Shared Function Create() As TesseractEngine
        ' TesseractEngine P/Invokes tesseract50.dll on Windows, libtesseract.so on Linux
        ' If the native binary for the current runtime identifier is missing, throws DllNotFoundException
        Return New TesseractEngine(ResolveTessDataPath(), "eng", EngineMode.Default)
    End Function
End Class

<ItemGroup>
  <None Update="tessdata\**\*">
    <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
  </None>
</ItemGroup>

// Install: dotnet add package IronOcr
//不是tessdata download.不是native binary configuration.不是path management.

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

var text = new IronTesseract().Read("document.jpg").Text;

' Install: dotnet add package IronOcr
' 不是tessdata download.不是native binary configuration.不是path management.

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"

Dim text As String = New IronTesseract().Read("document.jpg").Text

// dotnet add package IronOcr.Languages.French
// dotnet add package IronOcr.Languages.German

var ocr = new IronTesseract();
ocr.Language = OcrLanguage.French;
ocr.AddSecondaryLanguage(OcrLanguage.German);
var result = ocr.Read("multilingual-document.jpg");

Imports IronOcr

Dim ocr As New IronTesseract()
ocr.Language = OcrLanguage.French
ocr.AddSecondaryLanguage(OcrLanguage.German)
Dim result = ocr.Read("multilingual-document.jpg")

// charlesw/Tesseract: what you get without preprocessing
public string ExtractText(string imagePath)
{
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img = Pix.LoadFromFile(imagePath);
    using var page = engine.Process(img);

    // On a poor-quality scan: 40-70% accuracy
    // On a skewed document: 60-80% accuracy
    // On a photo of a document: 10-40% accuracy
    return page.GetText();
}

Imports Tesseract

Public Function ExtractText(imagePath As String) As String
    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(imagePath)
            Using page As Page = engine.Process(img)
                ' On a poor-quality scan: 40-70% accuracy
                ' On a skewed document: 60-80% accuracy
                ' On a photo of a document: 10-40% accuracy
                Return page.GetText()
            End Using
        End Using
    End Using
End Function

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

// Named preprocessing steps — no custom image processing code
input.Deskew();               // Corrects rotation up to ~10 degrees
input.DeNoise();              // Removes speckle and scanner noise
input.Contrast();             // Normalizes contrast
input.Binarize();             // Adaptive thresholding
input.EnhanceResolution(300); // Upsamples to 300 DPI if needed

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

Imports IronOcr

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

    ' Named preprocessing steps — no custom image processing code
    input.Deskew()               ' Corrects rotation up to ~10 degrees
    input.DeNoise()              ' Removes speckle and scanner noise
    input.Contrast()             ' Normalizes contrast
    input.Binarize()             ' Adaptive thresholding
    input.EnhanceResolution(300) ' Upsamples to 300 DPI if needed

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

// Platform-conditional code required to deploy charlesw/Tesseract
// This block exists in real production codebases

public static class TesseractFactory
{
    public static string GetTessDataPath()
    {
        // Runtime environment detection — purely deployment plumbing
        if (Environment.GetEnvironmentVariable("DOTNET_RUNNING_IN_CONTAINER") == "true")
        {
            // Docker: tessdata must be explicitly COPY'd into the image
            return "/app/tessdata";
        }

        if (RuntimeInformation.IsOSPlatform(OSPlatform.Linux))
        {
            // Linux bare metal: path convention differs from Windows
            return Path.Combine(AppContext.BaseDirectory, "tessdata");
        }

        // Windows dev machine
        return @"./tessdata";
    }

    public static TesseractEngine CreateEngine()
    {
        // x86/x64 conditional logic may be needed for specific deployment targets
        // EngineMode.Default uses LSTM; EngineMode.TesseractOnly uses legacy engine
        return new TesseractEngine(GetTessDataPath(), "eng", EngineMode.Default);
    }
}

Imports System
Imports System.IO
Imports System.Runtime.InteropServices

Public Module TesseractFactory
    Public Function GetTessDataPath() As String
        ' Runtime environment detection — purely deployment plumbing
        If Environment.GetEnvironmentVariable("DOTNET_RUNNING_IN_CONTAINER") = "true" Then
            ' Docker: tessdata must be explicitly COPY'd into the image
            Return "/app/tessdata"
        End If

        If RuntimeInformation.IsOSPlatform(OSPlatform.Linux) Then
            ' Linux bare metal: path convention differs from Windows
            Return Path.Combine(AppContext.BaseDirectory, "tessdata")
        End If

        ' Windows dev machine
        Return "./tessdata"
    End Function

    Public Function CreateEngine() As TesseractEngine
        ' x86/x64 conditional logic may be needed for specific deployment targets
        ' EngineMode.Default uses LSTM; EngineMode.TesseractOnly uses legacy engine
        Return New TesseractEngine(GetTessDataPath(), "eng", EngineMode.Default)
    End Function
End Module

// Before: charlesw/Tesseract
using var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default);
using var img = Pix.LoadFromFile(imagePath);
using var page = engine.Process(img);
var text = page.GetText();

// After: IronOCR
var text = new IronTesseract().Read(imagePath).Text;

Imports Tesseract

Dim text As String

Using engine As New TesseractEngine("./tessdata", "eng", EngineMode.Default)
    Using img As Pix = Pix.LoadFromFile(imagePath)
        Using page As Page = engine.Process(img)
            text = page.GetText()
        End Using
    End Using
End Using

' After: IronOCR
text = New IronTesseract().Read(imagePath).Text

// Before: charlesw/Tesseract iterator pattern
using var iter = page.GetIterator();
iter.Begin();
do
{
    if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
    {
        var word = iter.GetText(PageIteratorLevel.Word);
        var confidence = iter.GetConfidence(PageIteratorLevel.Word);
        Console.WriteLine($"'{word?.Trim()}' at ({bounds.X1},{bounds.Y1}) - {confidence:P1}");
    }
} while (iter.Next(PageIteratorLevel.Word));

// After:IronOCRdirect collection access
var result = new IronTesseract().Read(imagePath);
foreach (var word in result.Words)
{
    Console.WriteLine($"'{word.Text}' at ({word.X},{word.Y}) - {word.Confidence}%");
}

Imports IronTesseract

' Before: charlesw/Tesseract iterator pattern
Using iter = page.GetIterator()
    iter.Begin()
    Do
        Dim bounds As Rectangle
        If iter.TryGetBoundingBox(PageIteratorLevel.Word, bounds) Then
            Dim word = iter.GetText(PageIteratorLevel.Word)
            Dim confidence = iter.GetConfidence(PageIteratorLevel.Word)
            Console.WriteLine($"'{word?.Trim()}' at ({bounds.X1},{bounds.Y1}) - {confidence:P1}")
        End If
    Loop While iter.Next(PageIteratorLevel.Word)
End Using

' After: IronOCR direct collection access
Dim result = New IronTesseract().Read(imagePath)
For Each word In result.Words
    Console.WriteLine($"'{word.Text}' at ({word.X},{word.Y}) - {word.Confidence}%")
Next

// Before: charlesw/Tesseract + PdfiumViewer (two libraries, two native dependencies)
using (var document = PdfDocument.Load(pdfPath))
using (var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default))
{
    for (int i = 0; i < document.PageCount; i++)
    {
        using var pageImage = document.Render(i, 300, 300, PdfRenderFlags.CorrectFromDpi);
        // Save temp file, load into Pix, process, delete temp file...
    }
}

// After:IronOCR(one library, native PDF support)
var text = new IronTesseract().Read(pdfPath).Text;

Imports PdfiumViewer
Imports Tesseract

Using document As PdfDocument = PdfDocument.Load(pdfPath)
    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        For i As Integer = 0 To document.PageCount - 1
            Using pageImage = document.Render(i, 300, 300, PdfRenderFlags.CorrectFromDpi)
                ' Save temp file, load into Pix, process, delete temp file...
            End Using
        Next
    End Using
End Using

' After:IronOCR(one library, native PDF support)
Dim text As String = New IronTesseract().Read(pdfPath).Text

# Remove manual tessdata files and .csproj CopyToOutputDirectory entries
# AddNuGet語言包instead:
dotnet add package IronOcr.Languages.French
dotnet add package IronOcr.Languages.German