How to read information from bank cheque using IronOcr?

Question

Hello,

I have a sample Cheque. I am trying to read the following

a) Branch Name

b) Date

c) Pay to the order of

D) Cheque No.

I am using IronOcr. What extra thing I need to do for getting the relevant info since I am not able to get the informations properly.

Or any other OCR SDK is there specific to this purpose...Please help

Thanks in advance.

Mohamed Abedallah · Answer

I am not familiar with IronOcr, but you can use LEADTOOLS which has the ability to retrieve information from scanned bank cheque. The code is very simple and it is similar to this: // Initialize the RasterCodecs class RasterCodecs codecs = new RasterCodecs(); // Initialize the BankCheckReader class BankCheckReader checkReader = new BankCheckReader(); // The bank cheque image string chequePath = Path.Combine(@ "C:\Users\Public\Documents\LEADTOOLS Images\BankCheck.jpg" ); RasterImage image = codecs.Load(chequePath); IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD, false ); ocrEngine.Startup(codecs, null , null , null ); // Assign Engine to Reader checkReader.OcrEngine = ocrEngine; // Process Image checkReader.ProcessImage(image); ocrEngine.Shutdown(); //Print the information foreach (var value in checkReader.Results) { LeadRect bounds = value.Value.Bounds; Console.WriteLine( string .Format( "Field Name : {0}" , value.Key)); Console.WriteLine( string .Format( "Field Value : {0}" , value.Value.Text)); Console.WriteLine( string .Format( "Field Bounds : {0},{0},{0},{0}" , bounds.X.ToString(), bounds.Y.ToString(), bounds.Width.ToString(), bounds.Height.ToString())); Console.WriteLine( "************************************" ); } You can also check the online demo here: http://demo.leadtools.com/JavaScript/BankCheckReader/

Mehreen Tahir · Answer

Cheques are usually hand written and everything that is hand written is difficult to be accepted in any way. You have a chance with known or similar fonts or else tesseract can be helpful. It is probably the most accurate open source OCR engine available. Also there is a possibility to train tesseract for the User fonts. You can use the dotnet wrapper for this purpose. Find it here. https://www.nuget.org/packages/Tesseract/ Here are some good resources for training the tesseract. http://www.joyofdata.de/blog/a-guide-on-ocr-with-tesseract-3-03/ https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract Hope that helps.