FIINDER: FPGA Intensified-Image Networked Detector w/ Embedded Recognition

The system goal is a network-enabled FPGA-enhanced system for self-contained low-power image-intensified recognition. The system will support reprogramming, through-the-network, of its real-time recognition, FPGA supported identification of many asset classes. This effort focuses on feasibility questions for remote face-recognition. Facial recognition at significant distances is non-trivial and often subtle. Securics' founder led the only major study of facial recognition beyond 30 meters. This project expands that research to improve basic understanding of the critical issues in long-distance identification, addresses a novel research approach to using intensified images for face-recognition and provides for the design of a low-power, lightweight, reprogrammable, embedded recognition system. This project will collect the first ever Intensified Imagery (II) face database, and benchmark multiple algorithms on visible, II and LWIR imagery. This will include cross-modality testing, where one modality is enrolled and another is used for recognition. This project will address other issues by extending Securics' patent pending theory predicting failure of biometric systems and improving them via perturbation enhancing recognition of non-cooperating subjects in long-distance surveillance-oriented biometric identification. The final objective is the core of the FIINDER system including embedded systems design and prototyping, analysis of algorithms suitability for FPGA implementation, power-budget analysis, and field-testing.