What if 49% of the meal you ordered at a restaurant tasted good, but 51% tasted awful?
What if you slept only 49% of the time you were in bed, and you spent the rest of the night tossing and turning?
What if only 49% of the roads you drive on were smooth, and the rest was riddled with potholes and bumps?
None of those scenarios are acceptable, and you would demand better. You would send your meal back, change pillows, and drive on a different road the next time.
So why would you accept such poor accuracy from OCR technology?
Optical Character Recognition (OCR) is much faster than typing, but no one should be content with such an awful rate.
What if we could get 99% data accuracy?
At BIS, we were curious how accurate our data and document capture platform, Grooper, was compared to other popular OCR and data capture solutions. So we put Grooper to the test.
Using the same documents in all three steps of the test, we quickly processed and validated the results using:
The results proved what we suspected. Grooper’s unique tools make these once very inaccurate results highly accurate.
Grooper’s OCR Synthesis, Grooper’s Image Processing and Standard OCR | Grooper’s OCR Synthesis and Standard OCR | OCR alone |
99.91% | 78.09% | 49.60% |
Other data capture solutions are using OCR alone, or OCR with expensive many third-party add ons, to achieve somewhere between 49-78% accuracy.
This means that 51-22% of the text that they capture is riddled with errors or is completely unintelligible.