高度な文書分析

using Tesseract;
using System.Drawing;
using System.Threading.Tasks;

// Step 1: Convert PDF page to PNG image (requires separate PDF library)
// This example assumes you've already converted the scanned PDF to an image
string imagePath = "document-scan.png";

// Step 2: Initialize Tesseract with language data files path
var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default);

// Step 3: Load the input image and process
var img = Pix.LoadFromFile(imagePath);
var page = engine.Process(img);

// Step 4: Extract the recognized text
string extractedText = page.GetText();
Console.WriteLine($"Confidence: {page.GetMeanConfidence()}");
Console.WriteLine(extractedText);

// Optional: Get detailed results with bounding boxes
using (var iter = page.GetIterator())
{
    iter.Begin();
    do
    {
        if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
        {
            var word = iter.GetText(PageIteratorLevel.Word);
            Console.WriteLine($"Word: {word} at {bounds}");
        }
    } while (iter.Next(PageIteratorLevel.Word));
}

// Clean up resources
page.Dispose();
img.Dispose();
engine.Dispose();

Imports Tesseract
Imports System.Drawing
Imports System.Threading.Tasks

' Step 1: Convert PDF page to PNG image (requires separate PDF library)
' This example assumes you've already converted the scanned PDF to an image
Dim imagePath As String = "document-scan.png"

' Step 2: Initialize Tesseract with language data files path
Dim engine As New TesseractEngine("./tessdata", "eng", EngineMode.Default)

' Step 3: Load the input image and process
Dim img As Pix = Pix.LoadFromFile(imagePath)
Dim page As Page = engine.Process(img)

' Step 4: Extract the recognized text
Dim extractedText As String = page.GetText()
Console.WriteLine($"Confidence: {page.GetMeanConfidence()}")
Console.WriteLine(extractedText)

' Optional: Get detailed results with bounding boxes
Using iter As ResultIterator = page.GetIterator()
    iter.Begin()
    Do
        Dim bounds As Rect
        If iter.TryGetBoundingBox(PageIteratorLevel.Word, bounds) Then
            Dim word As String = iter.GetText(PageIteratorLevel.Word)
            Console.WriteLine($"Word: {word} at {bounds}")
        End If
    Loop While iter.Next(PageIteratorLevel.Word)
End Using

' Clean up resources
page.Dispose()
img.Dispose()
engine.Dispose()

// Example: Processing multiple PDF pages (after conversion)
public async Task<string> ProcessMultiPagePdf(string[] imagePaths)
{
    var results = new StringBuilder();
    var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default);

    foreach (var imagePath in imagePaths)
    {
        using (var img = Pix.LoadFromFile(imagePath))
        using (var page = engine.Process(img))
        {
            results.AppendLine($"Page confidence: {page.GetMeanConfidence():F2}");
            results.AppendLine(page.GetText());
            results.AppendLine("---");
        }
    }

    engine.Dispose();
    return results.ToString();
}

Imports System.Text
Imports Tesseract

Public Async Function ProcessMultiPagePdf(imagePaths As String()) As Task(Of String)
    Dim results As New StringBuilder()
    Dim engine As New TesseractEngine("./tessdata", "eng", EngineMode.Default)

    For Each imagePath In imagePaths
        Using img = Pix.LoadFromFile(imagePath)
            Using page = engine.Process(img)
                results.AppendLine($"Page confidence: {page.GetMeanConfidence():F2}")
                results.AppendLine(page.GetText())
                results.AppendLine("---")
            End Using
        End Using
    Next

    engine.Dispose()
    Return results.ToString()
End Function

using IronOcr;
using System.Linq;

// Initialize the OCR engine (improved Tesseract 5)
var ocr = new IronTesseract();

// Configure for improved accuracy
ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.Auto;
ocr.Configuration.ReadBarCodes = true; // Also detect barcodes/QR codes

// Load PDF document directly - no conversion needed
var input = new OcrInput();
input.LoadPdf("scanned-document.pdf", Password: "optional-password");

// Optional: Pre-process for better accuracy on low-quality scans
input.DeNoise();  // Remove noise from scanned paper documents
input.Deskew();   // Fix rotation from images captured at angles
input.EnhanceResolution(300); // Ensure improved DPI

// Extract text from all pages and create searchable data
OcrResult result = ocr.Read(input);

// Access detailed results
Console.WriteLine($"Overall Confidence: {result.Confidence}%");
Console.WriteLine($"Pages Processed: {result.Pages.Count()}");
Console.WriteLine(result.Text);

// Export as searchable PDF
result.SaveAsSearchablePdf("searchable-output.pdf");

Imports IronOcr
Imports System.Linq

' Initialize the OCR engine (improved Tesseract 5)
Dim ocr As New IronTesseract()

' Configure for improved accuracy
ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.Auto
ocr.Configuration.ReadBarCodes = True ' Also detect barcodes/QR codes

' Load PDF document directly - no conversion needed
Dim input As New OcrInput()
input.LoadPdf("scanned-document.pdf", Password:="optional-password")

' Optional: Pre-process for better accuracy on low-quality scans
input.DeNoise()  ' Remove noise from scanned paper documents
input.Deskew()   ' Fix rotation from images captured at angles
input.EnhanceResolution(300) ' Ensure improved DPI

' Extract text from all pages and create searchable data
Dim result As OcrResult = ocr.Read(input)

' Access detailed results
Console.WriteLine($"Overall Confidence: {result.Confidence}%")
Console.WriteLine($"Pages Processed: {result.Pages.Count()}")
Console.WriteLine(result.Text)

' Export as searchable PDF
result.SaveAsSearchablePdf("searchable-output.pdf")

using IronOcr;
using System.Drawing;

var ocr = new IronTesseract();

// Advanced configuration for specific document types
ocr.Configuration = new TesseractConfiguration()
{
    WhiteListCharacters = "0123456789.$,", // For financial documents
    BlackListCharacters = "`~", 
    PageSegmentationMode = TesseractPageSegmentationMode.SingleColumn
};

// Load specific pages from a PDF file (pages 1 and 2)
var input = new OcrInput();
input.LoadPdfPages("web-report.pdf", new[] { 0, 1 });

// Target specific regions for extraction (e.g., invoice totals)
var cropRegion = new CropRectangle(x: 100, y: 500, width: 400, height: 200);
foreach (var page in input.Pages)
{
    page.AddCropRegion(cropRegion);
}

// Perform OCR and get searchable text
OcrResult result = ocr.Read(input);

// Access structured data
foreach (var page in result.Pages)
{
    Console.WriteLine($"Page {page.PageNumber}:");
    foreach (var paragraph in page.Paragraphs)
    {
        Console.WriteLine($"  Paragraph (Confidence: {paragraph.Confidence}%):");
        Console.WriteLine($"  {paragraph.Text}");
    }
}

Imports IronOcr
Imports System.Drawing

Dim ocr = New IronTesseract()

' Advanced configuration for specific document types
ocr.Configuration = New TesseractConfiguration() With {
    .WhiteListCharacters = "0123456789.$,", ' For financial documents
    .BlackListCharacters = "`~",
    .PageSegmentationMode = TesseractPageSegmentationMode.SingleColumn
}

' Load specific pages from a PDF file (pages 1 and 2)
Dim input = New OcrInput()
input.LoadPdfPages("web-report.pdf", {0, 1})

' Target specific regions for extraction (e.g., invoice totals)
Dim cropRegion = New CropRectangle(x:=100, y:=500, width:=400, height:=200)
For Each page In input.Pages
    page.AddCropRegion(cropRegion)
Next

' Perform OCR and get searchable text
Dim result As OcrResult = ocr.Read(input)

' Access structured data
For Each page In result.Pages
    Console.WriteLine($"Page {page.PageNumber}:")
    For Each paragraph In page.Paragraphs
        Console.WriteLine($"  Paragraph (Confidence: {paragraph.Confidence}%):")
        Console.WriteLine($"  {paragraph.Text}")
    Next
Next