C#: OCR (Optical Character Recognition)

May 3, 2010 by Christoff Truter C#  

The past few weeks we've been looking for a suitable OCR solution to integrate into our document management system.

One option we came across involves MODI (Microsoft Office Document Imaging) - a tool available within Microsoft Office 2003 - 2007 (not available in Microsoft Office 2010).

Simply include the MODI Type library (COM Interop) and convert image(s) to text like this:

using MODI;
using System;

class Program
{
    static void Main(string[] args)
    {
        DocumentClass doc = new DocumentClass();
        doc.Create(@"some.tiff");
        doc.OCR(MiLANGUAGES.miLANG_ENGLISH, true, true);

        foreach (Image image in doc.Images)
        {
            Console.WriteLine(image.Layout.Text);
        }
    }
}

Its quite a powerful OCR engine, but the engine behind MODI isn't microsoft based - it is licensed under ScanSoft inc - currently Nuance.

There is one part I do find a bit dodgy though, we found quite a few rather expensive OCR tools out there (from $600), that integrates with MODI - which obviously requires Microsoft Office.

I almost feel that those application belong in the freeware realm - since you already bought a license to the core OCR functionality (via MS Office) and most of the non-OCR (part you will be paying for) seems rather mediocre.

My personal opinion though... ;)