OCR for recognising handwriting in .NET [closed]

问题:

I need to use an OCR component within a .NET application that can recognise handwriting. We got the MODI component from MS office up and running, and it recognises printed text, but not handwriting.

I don't mind paying.

Does anyone have any recommendations?

Many thanks Jim

回答1:

I recommend Pegasus Imaging, we use it every day and it is great. For handwriting you have the ICR module.

回答2:

In my application, we had a need to recognize some forms that were filled in with hand-printed text. The accuracy was good as long as the users used well-formed characters. The SDK I used is called Leadtools. My application uses MFC but they also have .NET classes.

回答3:

Most OCR toolkits are aimed at recognizing printed text, not handwritten text. I would recommend Atalasoft's toolkit for flexibility in that you can plug-in various OCR engines.

Recognizing handwritten text is much less accurate and more difficult. It is usually labelled as Intelligent Character Recognition (ICR), and I am not sure whether Atalasoft, Pegasus, or most commercial products support ICR. Hand-written recognition is aided by writing each letter in a cell on a grid, like filling out your name on the SAT. Are you talking about free form text or writing on a grid?

回答4:

Try tessnet, it's an open-source .NET OCR engine based on tesseract

回答5:

It's important to distinguish between hand-printed text and cursive script writing. The term ICR generally is used to refer to hand-printed text.

Accusoft Pegasus provides a SmartZone ICR .NET SDK that can get you recognizing text very quickly. Accuracy get better and better as you gain control over how the text is entered, such as on a form, etc.

It also is important to realize that the cues available to real-time recognition, such as on a tablet PC, are completely different than those available in a scanned image. This is why cursive recognition is so much easier in real-time than after the fact (scanned image). You can read the movements of the pen, and the order of strokes is very helpful in recognition.