A Framework for Evaluating OCR Engines’ Accuracy

Laserfiche Computer Science, 2019-20

Liaison(s): Tessa Adair ’14, Doren Lan ’18
Advisor(s): Xanda Schofield ’13
Students(s): Jocelyn Chen (PM-S), Athena Paraskevas-Nevius, Ian Taylor (PM-F), Andrew Lewis

Laserfiche specializes in software that enables organizations to replace manual, paper-based processes with digital records management, AI-powered workflows, e-forms and analytics. A critical component of this process is transferring existing paper documents onto the digital platform, where optical character recognition (OCR) is used to extract relevant information. To better inform Laserfiche about which OCR engines are best to use, our team developed metrics and a framework to test and compare the accuracy and performance of different available OCR engines.