Improving Unsatisfactory OCR Results

1. Start with ReadDocumentAdvanced() and EnhanceResolution()

Before reaching for any other configuration, pair the ReadDocumentAdvanced() method with the EnhanceResolution() preprocessing filter. This combination resolves most of the accuracy problems users hit with the standard read.

Apply the filter to the OcrInput, then pass that same input to ReadDocumentAdvanced():

var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadPdf("document.pdf");
// Upscale and sharpen low-resolution input before reading
input.EnhanceResolution();
var result = ocr.ReadDocumentAdvanced(input);
Console.WriteLine(result.Text);

var ocr = new IronTesseract();
using var input = new OcrInput();
input.LoadPdf("document.pdf");
// Upscale and sharpen low-resolution input before reading
input.EnhanceResolution();
var result = ocr.ReadDocumentAdvanced(input);
Console.WriteLine(result.Text);

Imports IronOcr

Dim ocr As New IronTesseract()
Using input As New OcrInput()
    input.LoadPdf("document.pdf")
    ' Upscale and sharpen low-resolution input before reading
    input.EnhanceResolution()
    Dim result = ocr.ReadDocumentAdvanced(input)
    Console.WriteLine(result.Text)
End Using

ReadDocumentAdvanced() runs the PaddleOCR machine-learning engine to parse layout-aware, text-heavy documents, while EnhanceResolution() upscales and sharpens low-resolution input before the read. Low resolution is a common root cause of poor results, so this preprocessing step often matters as much as the engine swap.

2. Target x64 on .NET Framework

Configure .NET Framework projects to build as x64. AdvancedScan needs a significant amount of memory, and running it under a different architecture can trigger runtime errors. See the Advanced Scan on .NET Framework guide for the full setup.