Which library to use to extract text from images?

Question

I am writing a program that when given an image of a low level math problem (e.g. 98*13) should be able to output the answer. The numbers would be black, and the background white. Not a captcha, just an image of a math problem.

The math problems would only have two numbers and one operator, and that operator would only be +, -, *, or /.

Obviously, I know how to do the calculating ;) I'm just not sure how to go about getting the text from the image.

A free library would be ideal... although If I have to write the code myself I could probably manage.

https://github.com/charlesw/tesseract/ "Getting started quickly" — camino, Apr 28 '19 at 00:55

score 8 · Answer 1 · answered Feb 28 '17 at 10:50

For extract words from image, I use the most accurate open source OCR engine: Tesseract. Available here or directly in your packages NuGet.

And this is my function in C#, which extract words from image passed in sourceFilePath. Set EngineMode to TesseractAndCube; it detect more word than the other options.

var path = "YourSolutionDirectoryPath";
using (var engine = new TesseractEngine(path + Path.DirectorySeparatorChar + "tessdata", "fra", EngineMode.TesseractAndCube))
{
    using (var img = Pix.LoadFromFile(sourceFilePath))
    {
        using (var page = engine.Process(img))
        {
            var text = page.GetText();
            // text variable contains a string with all words found
        }
    }
}

I hope that helps.

score 4 · Answer 2 · edited May 23 '17 at 12:32

4

Try this post regarding using the C++ Google Tessaract OCR lib in C#

OCR with the Tesseract interface

edited May 23 '17 at 12:32

Community

1
1

answered Feb 28 '11 at 19:30

Taylor Bird

7,767
1
25
31

score 2 · Answer 3 · answered Aug 19 '16 at 19:08

you can use Microsoft Office Document Imaging (Interop.MODI.dll) in visaul studio and extract text of pictures

Document modiDocument = new Document();
modiDocument.Create(filePath);
modiDocument.OCR(MiLANGUAGES.miLANG_ENGLISH);
MODI.Image modiImage = (modiDocument.Images[0] as MODI.Image);
string extractedText = modiImage.Layout.Text;
modiDocument.Close();
return extractedText;

score 2 · Answer 4 · answered Feb 28 '11 at 19:33

You need OCR. There is the free Tesseract library from Google, but it's C code. You could use in a C++/CLI project and access via .NET.

This article gives some information on recognizing numbers (for Sudoku, but your problem is similar)

http://sudokugrab.blogspot.com/2009/07/how-does-it-all-work.html

score 1 · Answer 5 · answered Jun 15 '22 at 05:04

IronOCR is free for development and testing. The default English language pack should do a good job of reading this, but you may also want to consider using a custom Tesseract language pack written specifically for equations.

See https://ironsoftware.com/csharp/ocr/languages/#custom-language-example

using IronOcr; 
var Ocr = new IronTesseract(); 
Ocr.UseCustomTesseractLanguageFile("languages/equ.traineddata");

using (var Input = new OcrInput(@"images\equation.png"))
{    
var Result = Ocr.Read(Input);
Console.WriteLine(Result.Text);
}

Disclaimer: I work for Iron Software.

Which library to use to extract text from images?

5 Answers5