Is Analytics and Machine Learning toolkit 2017 suitable for OCR?

[Di_Ne]

Hi all and NI Support especially,

A coworker wrote a program for OCR using Vision module. It works, but the more we train it, the easier it crashes. When it works, it gives a lot of misreadings.

Question : Is Analytics and Machine Learning toolkit 2017 suitable for OCR? We need to classify 36 classes : all upper case letters and numbers 0-9.

Thank you,

Di

P.S. : this board won't let me post unless I attach something

[Bob_Schor]

---

@Di_Ne

P.S. : this board won't let me post unless I attach something

---

That's strange -- I've never seen this requirement before, and seem to see lots of posts where someone asks "Why does my program have an error" without posting any code, any pictures, anything that helps us answer ...

Bob Schor

[RandallCrandall'sCandleHandle]

Hey Di_Ne,

I don't believe that this should crash all the time.

Are you programmatically training like in this example?

https://forums.ni.com/t5/Example-Programs/Programmatically-Train-IMAQ-OCR-Characters-using-LabVIEW/t...

In addition, any information you can give about the crash would help, because there are a lot of reasons that a program written in LabVIEW could crash. Do you get an error message with the crash? If so, a screenshot would probably be helpful.

Cheers,

Ryan C.