Ephesoft clearly unveils mobile document capture solution
The latest version of the technology takes advantage of innovations in everything from deep learning to mobile architecture to give consumers more accuracy and flexibility when it comes to document capture with their mobile devices.
Ike Kavas, Ephesoft CTO, highlighted the way the enhancements paved the way for better, easier-to-read images.
“For example, deep learning is employed in our live-edge detection functionality that has historically been a challenge for identifying document boundaries on mobile devices,” Kavas explains. “This new approach captures clearer images.”
Deep learning provides a better solution than typical bitonal image recognition because deep learning has predictive capabilities. The system is able to predict edges, for example, which along with the ability to extract text and numbers, makes it ideal for credit card recognition.
Ephesoft’s Transact Mobile SDK 4.0 also includes image auto-capture and auto-alignment, making images scanned by mobile devices comparable to those created by topline scanners.
Online and offline batch processing enables users to do character recognition, document classification, and data extraction when offline. Version 4.0 also features security enhancements such as a private directory for storing images temporarily when using mobile devices (the directory is erased when the application is closed).
The solution uses ARM 64-bit architecture to improve memory capacity for more powerful processing, and also includes image enhancement tools, simple barcode interpretation, and on-device OCR.
Founded in 2010 and headquartered in Laguna Hills, California, Ephesoft has raised $15 million in funding, completing a Series A round led by Mercato Partners in August.