I've noticed the OCR is unable to detect a single character. I read in another thread from last year: https://forum.ocrsdk.com/thread/why-real-time-recognition-sdk-cannot-recognize-a-single-word/?order=all#comment-e554448a-bac3-47ff-82e3-a7df00b3a4d9 That this was expected to have been resolved by the end of last year.
Am i using an older version of the SDK or has this not yet been resolved?
If so is there any expectations on when this would be resolved?
Single Character Recognition
Was this article helpful?
0 out of 0 found this helpful
Comments
7 comments
Please try to use our standard data capture scenario that is available in the Real-Time Recognition (sample-datacapture). You can set up a regular expression for the separate field to extract a single character. More information about this scenario please find in the following articles:
Capture a Custom Data Field with iOS for iOS
Capture a Custom Data Field with Android for Android
Hi Diana,
Thanks for getting back to me.
This unfortunately would not suit our use case as this would only allow the capture of one field at at time.
The application using RTR SDK requires a column of numbers to be captured in real time, including single characters.
This would require the text capture scenario to be able to capture single characters OR the data capture to be able to capture multiple fields of different types of expressions in one specified area of interest in the same way the text capture scenario does.
Is this something expected to be implemented into the SDK?
According to our analysts the current version of RTR doesn't provide functionality to detect single characters, but at the moment we are working on improving this recognition scenario.
Thanks for letting me know Diana,
At this stage is it possible to say when this is expected to be rolled out ?
Unfortunately, we can not specify the exact dates. Perhaps, we will have this information in September.
Hi Diana, do you have any more information on this functionality yet?
I also require a mix of single and multiple characters (numbers in my case)
Hi Paul,
The team is continuously working on the enhancements of the text detection algorithm. Currently, detection of single characters, as well as improvements CJK text detection, are in focus. We hope to have the next update in Q2 2019.
Please sign in to leave a comment.