Digital Check was the first company to provide a MICR optical character recognition (OCR) verification capability for desktop check scanners, introducing the technology in 2002. Multiple magnetic MICR read routines work together with OCR verification, using custom algorithms to identify, evaluate and deliver the highest confidence MICR read rates and virtually eliminate miss-reads.
“MICR results are best measured in ‘real world’ environments, and we rely upon our solution partners and bank customers to provide on-going feedback to continually enhance our API and optimize our MICR performance. Our customers have told us repeatedly that our multiple MICR reads with OCR verification consistently deliver between 99% to virtually 100% accuracy and can be as much as three to four points better than other approaches,” said John Gainer, executive vice president of Digital Check. “More importantly, these same customers stress that the cost to correct a low confidence character by an operator at the time of check scanning runs from 25 to 30 cents, and they estimate that the cost of correcting an erroneous MICR character within downstream operations is $6 to $9 dollars per transaction. As the number of checks captured by remote check scanners accelerates, even incremental improvements in our MICR OCR verification capabilities will translate into large savings for financial institutions that have widespread deployments of remote deposit capture and branch capture applications.”
Digital Check’s Best Read Image technology uses adaptive thresholding logic to create cleaner check images and enable higher capture of critical check data by software applications. CAR/LAR engines that read the courtesy and legal amount fields require images that have clean backgrounds to achieve the best confidence levels and optimum read rates. Many desktop scanners use a fixed thresholding method to produce a bitonal (black and white) image, which can inhibit optimal character recognition. Using adaptive thresholding, the Digital Check scanner runs multiple thresholds to remove unwanted background “noise” until it selects the image that produces the best quality for CAR/LAR recognition.
This ability to re-threshold at the time the document is scanned occurs within the Digital Check API and represents a critical success factor in the remote deposit capture software application’s ability to optimize CAR/LAR read rates without operator intervention.
All Digital Check TellerScan models use the common API, which also features USB 2.0 plug-and-play connections. This technology makes adding or upgrading a check scanner device a quick and easy process for users.
