Saturday, November 23, 2013

A Genealogy Dream in the Making: OCR Software That Reads Old Handwriting

Diane Haddad on the Genealogy Insider blog is reporting that "Mocavo announced this week that it's making progress on optical character recognition (OCR) software that will read cursive handwriting, which could revolutionize how digitized records are put online."

On the Mocavo blog, company founder Cliff Shaw described the process, which first involved developing OCR software that could  "perfectly separate handwriting from typewritten text."

Now, Shaw says, the company is getting closer to the "Holy Grail" of being able to accurately read handwritten text. "With limited vocabularies (potential answers), we’re achieving 90-95% accuracy," he writes.

They still have work to do to achieve the ability to read handwriting of a wide range of styles, and to overcome problems with faded or ink-spotted documents I mentioned above. Read about the software and see examples on the Mocavo blog.

Imagine having  software package that can index a hand written document such as this without human intervention. That would truly be the holy grail of software.