This may be hard to explain but I'm looking for any alternative solutions. We have a client that plans on taking a few thousand of these screengrabs (see attached) and OCRing them to better search by serial, description, etc. For simplicity's sake, the highlighted "ARM-CLUTCH PEDAL" will be OCR'd and everything is good so far. The problem comes in when we take the second screen grab. The second screen grab will look almost the same as the first but the picture will be different and the highlighted item, which would now be "CROSS SHAFT-CLUTCH CONT" now needs to be OCR'd instead. Is there any way that quick fields can pick up what area is highlighted to OCR that section? Let me know if you need any other clarity.
Question
Question
Auto-Shifting Zone OCR
asked on June 2, 2016
0
0
Replies
replied on June 2, 2016
Does the number at the bottom always match the highlighted rows?
0
0
replied on June 2, 2016
Yes. However there are some lines that coincide with the above number. (see another attached example.)
0
0
replied on June 7, 2016
How are the files named? Can you use pattern matching to pull anything out of the name?
0
0
replied on June 8, 2016
That was another one of my thoughts, however, this old system that we are screen grabbing for is incredibly inconsistent in its naming conventions and even some reference numbers beside it aren't even there.
However, this does pose a side question. With Pattern Matching, am I able to tell it to identify anywhere on the page that is a three digit numerical value and OCR the information to the right/left of that number?
0
0
You are not allowed to follow up in this post.