First American Financial Corporation, a global provider of title insurance, settlement services and risk solutions for real estate transactions, has been issued a patent for an innovative improvement to optical character recognition (OCR) technology that leverages artificial intelligence.

The patented enhancement applies artificial intelligence to rapidly, efficiently, and accurately validate and correct text obtained from a document image via OCR, fuelling more accurate automated title underwriting and production on key text-based components of the title policy.

"The new patent for OCR Error Correction is the eighth patent we have been awarded in the last 12 months and further demonstrates First American’s continuing leadership in driving innovation within the title and settlement services industry," said Dennis J. Gilmore, CEO, First American Financial Corporation.

"First American has been at the forefront of title and settlement innovation for decades, leading the industry shift to computer-based title plants in the late 20^th century, and continues to drive innovation today, leveraging artificial intelligence and machine learning technologies to accelerate the digital transformation of the real estate settlement process."

The U.S. Patent No. 10,896,292 titled "OCR Error Correction" is the 30^th active patent in First American’s portfolio, which includes:

• Patents on technologies that facilitate the automated underwriting and production of title insurance;
• Patents on technologies that enable risk assessments of mortgage loans;
• Patents on technologies that improve the speed and precision of data collection or extraction from source documents, including documents essential to the underwriting of title insurance policies;
• Patents related to optical character recognition technologies that help to unlock and structure the data held in millions of public records, powering automated title production breakthroughs, as well as enhancing the breadth and accuracy of title plant data.

"Automated title production and accelerating the digital transformation of real estate transactions depend upon obtaining accurate, structured data from document images. This patented method allows us to transform unstructured data pulled from public records and other sources into structured data more rapidly and with greater accuracy than conventional OCR validation and error correction techniques," said Mark Fleming, senior vice president of Decision Sciences at First American.

"This patent builds on our rich history of innovation, including pioneering a variety of OCR technology advancements and other data extraction techniques that help move our automated title underwriting and production efforts forward."