Concentric AI launches vision-model feature for sensitive document discovery
Concentric AI said it made a vision-model feature available in its Concentric AI Semantic Intelligence platform to detect sensitive documents with unique visual signatures, including passports and driver licenses. The company linked the capability to Sensitive Data Discovery for documents where visual characteristics can be used to identify document type.
Concentric AI said organizations face challenges discovering sensitive data across environments in various formats. It said visual information in a data record may provide more signal than text, and the feature analyzes an image to identify document type based on visually consistent characteristics, including when contents are blurred.
The feature uses image content analysis to understand the biographic (or photo) page of a passport from any country and a driver license from any state in the U.S. Concentric AI contrasted the approach with optical character recognition, which it said requires processing every pixel in an image to detect and identify characters for machine-readable text and may struggle with low-quality images.
Concentric AI said the new feature provides an alternative to OCR limitations and that partner models can be built for effectively discovering sensitive data based on visually consistent characteristics in partners’ customers’ unique environments. “Many sensitive documents carry a distinct visual identity,” said Dr. Madhu Shashanka, Co-Chief Technology Officer and Co-Founder, Concentric AI.
The original article was written by Decision Insights Editorial.