Character Set File for OCR

[Epons]

Hi everybody,

I'm a new user from Labview, well, I'm new in a lot of topics....

I need to read a printed document (like Arial font or similar). The question is about character set available in vision builder. I've found three files, OCRA20, OCRA30 and SEMIDOT but they aren't sufficient to "read" an easy text. I think that the problem is about character set, so I need to train the OCR. Does someone know if exist a place to get a trained file for "standard" fonts? I don't know if the OCR strongly depends of distance, or others parameters and I need to train OCR for each application.

Thank you very much.

Regards,

Enric

[Minh Pham]

I would think you had better to create your own Character Set for your application, every application is different in many different ways. Talking about the Character Recognition, each font type has different set of characters itself. So i believe you would need to train the system as much as you can so that the character set is well recognized by the System.

VIsion is all about images so all of the important factors like clear, brightness, contrast, distance taken etc... need to take into account while creating your character set, the better you train your system, the better its intelligent in recognizing things. Just like us, without training ourselves we wouldnt be able to recognise many different objects.

The training tool in Vision quite simple to use, play around with parrametters while creating your Character Set to achieve better results.

Good luck

[Epons]

Hi Minh,

thank you for your answer. I will start to acquire samples to train the system and get a good Character Set for my application.

I agree with you about, without training we wouldn't be able to recognise objects or others things nevertheless we are training forever, everyday, for the rest of our lifes. There are a lot of things to know... (It's a little philosophical observation).

Thank you again.

Enric .....training my brain and my application.....