See comparisons of Traditional OCR vs. Grooper AI-Based OCR
Let's see it
OCR for Check Extraction
T
Tim McMullin
Updated: Apr 04, 2025
Description
A quick demo highlighting Grooper's combined OCR and handwriting recognition capabilities.
Steps
1
Watch a Demo of Grooper's OCR!
2
There are
3
We'll start our demo in the Grooper Design Studio. Here we can test all things Grooper. In this case, we're going to test OCR profiles.
4
Check processing used to be difficult. There are three challenges here - machine-printed text, handwriting, and a MICR line. There used to be specialized software just for checks.
5
Grooper makes quick work of the handwriting.
6
Let's show a traditional OCR profile - one of Grooper's best when there's no handwriting invovled.
7
Drastically different results. All this engine gets is the machine printing - notice the handwriting and MICR are unusable.
8
This view is of the raw text that comes back from the OCR engine. Again, notice how bad it is. Take special notice of "garbage" characters.
9
We'll re-run the document with the handwriting profile so you can see the difference live.
10
Grooper creates a synthetic view that uses the location information to make an approximate image from the extracted data... looks pretty good. Now let's look at the text.
11
The text is perfect. This is important because clean text is needed for AI downstream.