Research Area: Statistical Evaluation of Error Rates for Graph-Based Identification Technologies

Project Description: Graph-based technologies have proven to be an effective method for dealing with the complexities of handwriting biometric data. These methods have demonstrated high accuracy rates in a number of languages. Preliminary research has also demonstrated that graph-based technologies can be applied to other biometric modalities besides handwriting. An important tool for extending graph-based methods to new modalities is the ability to provide a statistical estimate of the accuracy of the procedure. The accuracy of the method will depend upon the quality of the quantification procedure and the ability of the statistical component of the classifier to use the information retained in the quantification procedure. The accuracy of a classification method or the corresponding error rates are parameters that can be statistically estimated. The primary activity of the proposed research is to develop a standard set of tools to estimate the accuracy of a classification procedure. This research will facilitate the development and extension of graph-based quantification technologies to new modalities, thereby providing concrete guidance on the improvements gained or not gained over traditional methods.

Requirements: A U.S. citizen who has recently graduated (or will be graduating by the fellowship start date) from an accredited Ph.D. program within the last five years and is interested in working as a Research Assistant Professor with the Intelligence and Security Research Center at George Mason University.

IC Postdoc Website: http://www.icpostdoc.org/

To apply: If you are interested in joining this research project as a Postdoctoral Fellow, please contact:
Dr. Christopher Saunders, Principal Investigator
George Mason University
Fairfax, Virginia 22030
703-993-9333
csaunde6@gmu.edu

The Document Forensics Laboratory is a joint venture by the Department of Applied Information Technology and Department of Statistics. Funded by the Gannon Technology Groups, LLC, the principal goal is to use a fusion of advanced statistical analysis techniques and information technology to determine the extent to which handwriting can be used as an accurate biometric identifier. Of particular interest are scripts other than Latin. The work of the laboratory has important consequences in national security.