01-03-2009, 10:47 AM
Automated Authentication of Identity Documents
Identity documents (IDs), such as passports and drivers' licenses are relied upon to deter fraud and stop terrorism. A multitude of document types and increased expertise in forgery make human inspection of such documents inconsistent and error prone. New generation reader/authenticator technology can assist in the ID screening process.
Such devices can read the information on the ID, authenticate it, and provide an overall security risk analysis. This talk will discuss how image processing and pattern recognition technology were used in the implementation of one such commercial device, the AssureTec i-Dentify reader.
The reader is based on a high resolution color CCD camera which automatically captures a presented ID under a variety of light sources (Visible, UV, IR, and others) in a few seconds.
Automated processing of IDs involves a number of interesting technical challenges which will be discussed: sensing the presence of a document in the reader
viewing area; cropping the document and extracting its size; identifying the document type by rapid comparison to a known document library; locating, extracting, and image processing of data fields of various types (text, photo, symbols, barcodes); processing text fields with appropriate OCR engines; cross-checking data from different parts of a document for consistence; checking for the presence of security features (e.g., UV patterns); and providing an overall risk assessment that the document is falsified. A live demonstration of the AssureTec i-Dentify reader will be given.