Azure OCR 與 IronOCR：哪種光學字元辨識解決方案最適合 .NET 專案？

// Windows.Media.Ocr: 6+ async steps before receiving any text
// Requires net8.0-windows10.0.19041.0 TFM — will not compile cross-platform

public async Task<string> ExtractTextAsync(string imagePath)
{
    // Step 1: WinRT file system access
    var file = await StorageFile.GetFileFromPathAsync(imagePath);

    // Step 2: Open WinRT stream
    using var stream = await file.OpenAsync(FileAccessMode.Read);

    // Step 3: Create bitmap decoder
    var decoder = await BitmapDecoder.CreateAsync(stream);

    // Step 4: Decode to SoftwareBitmap
    var bitmap = await decoder.GetSoftwareBitmapAsync();

    // Step 5: Check language availability — null if not installed on this machine
    var engine = OcrEngine.TryCreateFromLanguage(
        new Windows.Globalization.Language("en-US"));
    if (engine == null)
        throw new Exception("OCR engine not available for this language");

    // Step 6: Recognize
    var result = await engine.RecognizeAsync(bitmap);
    return result.Text;
}

Imports System.Threading.Tasks
Imports Windows.Media.Ocr
Imports Windows.Storage
Imports Windows.Graphics.Imaging
Imports Windows.Storage.Streams
Imports Windows.Globalization

Public Async Function ExtractTextAsync(imagePath As String) As Task(Of String)
    ' Step 1: WinRT file system access
    Dim file As StorageFile = Await StorageFile.GetFileFromPathAsync(imagePath)

    ' Step 2: Open WinRT stream
    Using stream As IRandomAccessStream = Await file.OpenAsync(FileAccessMode.Read)

        ' Step 3: Create bitmap decoder
        Dim decoder As BitmapDecoder = Await BitmapDecoder.CreateAsync(stream)

        ' Step 4: Decode to SoftwareBitmap
        Dim bitmap As SoftwareBitmap = Await decoder.GetSoftwareBitmapAsync()

        ' Step 5: Check language availability — Nothing if not installed on this machine
        Dim engine As OcrEngine = OcrEngine.TryCreateFromLanguage(New Language("en-US"))
        If engine Is Nothing Then
            Throw New Exception("OCR engine not available for this language")
        End If

        ' Step 6: Recognize
        Dim result As OcrResult = Await engine.RecognizeAsync(bitmap)
        Return result.Text
    End Using
End Function

<PropertyGroup>
  <TargetFramework>net8.0-windows10.0.19041.0</TargetFramework>
</PropertyGroup>

// Same code — Windows, Linux, macOS, Docker, AWS Lambda
//不是platform TFM, no WinRT, no conditional compilation
using IronOcr;

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";
var text = new IronTesseract().Read("document.jpg").Text;

Imports IronOcr

' Same code — Windows, Linux, macOS, Docker, AWS Lambda
'不是platform TFM, no WinRT, no conditional compilation

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"
Dim text As String = New IronTesseract().Read("document.jpg").Text

// Windows.Media.Ocr: language availability is a runtime unknown
// Returns null if the language pack is not installed on this machine

var engine = OcrEngine.TryCreateFromLanguage(
    new Windows.Globalization.Language("fr-FR"));

if (engine == null)
{
    // French OCR is simply unavailable — no recovery path
    // User must go to Windows Settings > Language to install French
    throw new InvalidOperationException(
        "French OCR unavailable. Install the French language pack in Windows Settings.");
}

var result = await engine.RecognizeAsync(bitmap);

Imports Windows.Globalization
Imports Windows.Media.Ocr

' Windows.Media.Ocr: language availability is a runtime unknown
' Returns Nothing if the language pack is not installed on this machine

Dim engine = OcrEngine.TryCreateFromLanguage(New Language("fr-FR"))

If engine Is Nothing Then
    ' French OCR is simply unavailable — no recovery path
    ' User must go to Windows Settings > Language to install French
    Throw New InvalidOperationException("French OCR unavailable. Install the French language pack in Windows Settings.")
End If

Dim result = Await engine.RecognizeAsync(bitmap)

// IronOCR: language availability is a deploy-time guarantee, not a runtime unknown
var ocr = new IronTesseract();
ocr.Language = OcrLanguage.French;
ocr.AddSecondaryLanguage(OcrLanguage.German);

// Works on any machine, any OS, zero OS configuration required
var result = ocr.Read("multilingual-document.jpg");
Console.WriteLine(result.Text);

Imports IronOcr

' IronOCR: language availability is a deploy-time guarantee, not a runtime unknown
Dim ocr As New IronTesseract()
ocr.Language = OcrLanguage.French
ocr.AddSecondaryLanguage(OcrLanguage.German)

' Works on any machine, any OS, zero OS configuration required
Dim result = ocr.Read("multilingual-document.jpg")
Console.WriteLine(result.Text)

// Windows.Media.Ocr: no preprocessing — what you pass is what gets recognized
// Skewed, noisy, or low-resolution images degrade accuracy with no remedy
// Manual preprocessing via separate Windows Imaging APIs is the only option

var bitmap = await decoder.GetSoftwareBitmapAsync();
// bitmap goes directly to recognition with no quality improvement
var result = await engine.RecognizeAsync(bitmap);

Imports System.Threading.Tasks

' Windows.Media.Ocr: no preprocessing — what you pass is what gets recognized
' Skewed, noisy, or low-resolution images degrade accuracy with no remedy
' Manual preprocessing via separate Windows Imaging APIs is the only option

Dim bitmap = Await decoder.GetSoftwareBitmapAsync()
' bitmap goes directly to recognition with no quality improvement
Dim result = Await engine.RecognizeAsync(bitmap)

// IronOCR: explicit preprocessing pipeline
// Each filter targets a specific quality defect
using var input = new OcrInput();
input.LoadImage("low-quality-scan.jpg");

input.Deskew();              // Correct page rotation up to ~40 degrees
input.DeNoise();             // Remove scanner speckle and compression artifacts
input.Contrast();            // Boost contrast on faded or washed-out text
input.Binarize();            // Convert to black/white with optimal threshold
input.EnhanceResolution(300); // Scale image to 300 DPI for recognition

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

Imports IronOcr

' IronOCR: explicit preprocessing pipeline
' Each filter targets a specific quality defect
Using input As New OcrInput()
    input.LoadImage("low-quality-scan.jpg")

    input.Deskew()              ' Correct page rotation up to ~40 degrees
    input.DeNoise()             ' Remove scanner speckle and compression artifacts
    input.Contrast()            ' Boost contrast on faded or washed-out text
    input.Binarize()            ' Convert to black/white with optimal threshold
    input.EnhanceResolution(300) ' Scale image to 300 DPI for recognition

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

// Windows.Media.Ocr: no PDF support
// Requires external PDF renderer to rasterize pages before OCR
// Conceptual pattern — a PDF rendering library is not provided by Windows APIs

// Step 1: Use external PDF library to render page to bitmap (not shown)
// Step 2: Pass rendered bitmap to Windows OCR
// var bitmap = RenderPdfPageToBitmap(pdfPath, pageIndex); // external library required

var engine = OcrEngine.TryCreateFromUserProfileLanguages();
if (engine == null)
    throw new Exception("No OCR language available");

// Step 3: Recognize the rasterized page
// var result = await engine.RecognizeAsync(bitmap);
// Step 4: Collect and concatenate results across all pages manually

' Windows.Media.Ocr: no PDF support
' Requires external PDF renderer to rasterize pages before OCR
' Conceptual pattern — a PDF rendering library is not provided by Windows APIs

' Step 1: Use external PDF library to render page to bitmap (not shown)
' Step 2: Pass rendered bitmap to Windows OCR
' Dim bitmap = RenderPdfPageToBitmap(pdfPath, pageIndex) ' external library required

Dim engine = OcrEngine.TryCreateFromUserProfileLanguages()
If engine Is Nothing Then
    Throw New Exception("No OCR language available")
End If

' Step 3: Recognize the rasterized page
' Dim result = Await engine.RecognizeAsync(bitmap)
' Step 4: Collect and concatenate results across all pages manually

// IronOCR: native PDF support — no external renderer needed
var text = new IronTesseract().Read("scanned-document.pdf").Text;

// Password-protected PDFs
using var input = new OcrInput();
input.LoadPdf("encrypted.pdf", Password: "secret");
var result = new IronTesseract().Read(input);
Console.WriteLine(result.Text);

//可搜尋的PDFoutput: make a scanned PDF text-searchable
var ocrResult = new IronTesseract().Read("scanned-archive.pdf");
ocrResult.SaveAsSearchablePdf("searchable-output.pdf");

Imports IronOcr

' IronOCR: native PDF support — no external renderer needed
Dim text As String = New IronTesseract().Read("scanned-document.pdf").Text

' Password-protected PDFs
Using input As New OcrInput()
    input.LoadPdf("encrypted.pdf", Password:="secret")
    Dim result = New IronTesseract().Read(input)
    Console.WriteLine(result.Text)
End Using

' 可搜尋的PDFoutput: make a scanned PDF text-searchable
Dim ocrResult = New IronTesseract().Read("scanned-archive.pdf")
ocrResult.SaveAsSearchablePdf("searchable-output.pdf")

// Before: Windows.Media.Ocr — 6+ await operations
var file = await StorageFile.GetFileFromPathAsync(imagePath);
using var stream = await file.OpenAsync(FileAccessMode.Read);
var decoder = await BitmapDecoder.CreateAsync(stream);
var bitmap = await decoder.GetSoftwareBitmapAsync();
var engine = OcrEngine.TryCreateFromUserProfileLanguages();
var winResult = await engine.RecognizeAsync(bitmap);
string text = winResult.Text;

// After:IronOCR— 1 call, same result, any platform
string text = new IronTesseract().Read(imagePath).Text;

Imports Windows.Storage
Imports Windows.Graphics.Imaging
Imports Windows.Media.Ocr

' Before: Windows.Media.Ocr — 6+ await operations
Dim file As StorageFile = Await StorageFile.GetFileFromPathAsync(imagePath)
Using stream = Await file.OpenAsync(FileAccessMode.Read)
    Dim decoder As BitmapDecoder = Await BitmapDecoder.CreateAsync(stream)
    Dim bitmap = Await decoder.GetSoftwareBitmapAsync()
    Dim engine = OcrEngine.TryCreateFromUserProfileLanguages()
    Dim winResult = Await engine.RecognizeAsync(bitmap)
    Dim text As String = winResult.Text
End Using

' After: IronOCR— 1 call, same result, any platform
Dim text As String = New IronTesseract().Read(imagePath).Text