SBIR-STTR Award

The use of contextual information is often a weak, error prone and labor intensive method of identifying and tracking Persons of Interest. Using the Neyman-Pearson Lemma, Ultra-Scan will fuse large numbers of historically weak contextual data fields to create accurate, high value identity information. The technical objective is to identify independent or weakly correlated contextual fields that can be treated as a score-based recognition system suitable for the Neyman Pearson Test, and which can then be used to significantly improve overall identification system performance. Phase I will research a large number of contextual personal identifiers that create an accurate form of personal identification when fused. The effort will create an ideal platform from which to implement a series of steps involving analysis, data modeling, estimation and software simulation to establish with mathematical certainty the ability to fuse large number of contextual fields to create a reliable form of identification.

Keywords:
Contextual Data Fusion, Neyman-Pearson, Multimodal Biometric Fusion, Identity Management, Poi-Person Of Interest

The use of contextual information is often a weak, error prone and labor intensive method of identifying and tracking Persons of Interest (POI). Ultra-Scan’s Phase I work successfully applied Neyman-Pearson fusion to weak contextual data, resulting in a powerful new algorithm that fuses discrete contextual data with biometrics in order to significantly improve system accuracy. Ultra-Scan’s algorithm actually increases performance as more weak data is fused. As a result of the Phase I effort, Ultra-Scan has established a solid foundation for the Phase II goal of developing a framework and methods for improving profile information and POI tracking, as well as constructing a prototype system to establish proof-of-concept. Upon completion of the Phase II effort, Ultra-Scan expects to deliver a prototype Interoperable Contextual Fusion Platform that will: Refine the company’s contextual fusion algorithm; Advance the fusion of contextual and biometric data for improved system accuracy; Create an automated, intuitive and customizable contextual mapping interface; Significantly improve the profiling of POIs for ID and tracking across disparate databases.

Keywords:
Contextual Data Fusion, Multimodal Biometric Fusion, Identity Management, Poi-Person Of Interest, Neyman-Pearson