PRESS RELEASES
Mark Walch, President of Gannon Technology Group, to present at the American Academy of Forensic Sciences Annual Meeting
(January 2008)
Alexandria, VA – January 21, 2008 - Gannon Technologies Group, the leader in graph based pattern recognition technology, announced today that it will speak and present a scientific whitepaper at the American Academy of Forensic Sciences annual meeting.
Mark Walch , President of Gannon Technologies Group, will speak on “FLASH ID: A Totally Automated, Language Independent Approach for Using Handwriting as a Biometric Identifier”. FLASH ID is the Federal Bureau of Investigation’s implementation of Gannon’s FLEX-Tracker handwriting biometric product.
This presentation will expose attendees to an innovative and highly effective means for automatic handwriting derived biometric identification. The centerpiece of the presentation will be the FLEX-Tracker software package developed by the Gannon Technologies Group in collaboration with George Mason University with extensive guidance and technical input provided by forensic scientists from the FBI Laboratory.
FLASH ID will be presented as a fully operational software system that can address the immediate needs within the forensic community related to using handwriting as a biometric identifier. The presenter will illustrate how individual features, available and quantifiable within an individual’s writing, can be empirically captured into a “loss less” data structure that preserves the topology and geometry of the original writing. The presentation will continue into the statistical analysis of this data structure to capture those elements that link the writing to its writer. A step-by-step demonstration will be provided showing the methods for taking known writing samples and capturing them as a data structure based on Graph Theory replete with both topology and hundreds of detailed physical measurements. It will then be shown how this data structure can be analyzed using statistical methods to distill the topological and physical features into a “biometric kernel”. The Biometric Kernel is the statistically derived subset of those measurements that truly captures the essence of an individual’s writing. Otherwise stated, the Biometric Kernel consists of those features that hold most consistent within an individual’s writing and vary the most across multiple different writers. Once the Biometric Kernel is established, FLASH ID can act on any unknown sample of handwriting and will return the nearest value in its handwriting reference database that provides the closest match to the questioned writing sample.
A key point to be made to the audience is FLASH ID represents a new approach toward using handwriting as a biometric identifier that does not attempt to replicate the actions of a Forensic Document Examiner. But rather, it brings to bear the power of what computers do very well—rapid capture and processing of large quantities of data—into the hands of forensic experts.
The core message will be rooted in two important aspects of the technology used to build FLASH ID. First, FLASH ID represents a totally automated process for extracting graphical data from handwritten documents, analyzing this data using established statistical methods and matching documents based on similarity of the captured writing. Second, the technology underlying FLASH ID is language independent. That is, the empirical and analytical techniques that power the handwriting derived biometric process have been demonstrated to function in different languages with completely different scripts.
As a residual biometric that can link individuals to documents, handwriting provides an important data source for both law enforcement and intelligence purposes. In the form of FLASH ID, the forensic science community will now have a tool that harnesses the power of automation to leverage the effectiveness of Document Examiners by capturing similarities embedded among multiple writing samples and graphically showcasing these similarities supported by the statistical analysis that led to their identification. FLASH ID will also extend document forensics across language barriers—something that is not commonly practiced today.
The Academy's annual scientific meeting is held in February at which time over 500 scientific papers, breakfast seminars, workshops, and other special events are presented. The AAFS consists of ten sections representing a wide range of forensic specialties, and the annual scientific meeting gathers these professionals who present the most current information, research, and updates in this expanding field.
The Academy's 60th Anniversary Scientific Meeting will be held February 18-23, 2008, at the Marriott Wardman Park Hotel in Washington, DC.
The American Academy of Forensic Sciences (AAFS), founded in 1948, is headquartered in Colorado Springs, CO. AAFS serves a distinguished and diverse membership consisting of nearly 6,000 members, divided into ten sections spanning the forensic enterprise. As a professional society dedicated to the application of science to the law, the AAFS is committed to the promotion of education and the elevation of accuracy, precision, and specificity in the forensic sciences. As the world’s most prestigious forensic science organization, the AAFS represents its membership to the public and serves as the focal point for public information concerning the forensic science profession.
Interested in learning more about GTG’s groundbreaking work?
Contact us at info@gannontech.