Self-Calibrating Multi-Mode Filter Spectroradiometer with Measurement Applications for Column Amounts of O3, NO2, SO2 and Aerosol Optical Depths

Not Available ARGTEC is pioneering the use of a new, innovative technology based on the theory of Attributed Relational Graph for content-based indexing and retrieval in large information databases. We have achieved several breakthroughs in the development of this technology. We applied this approach to indexing and retrieval in Automated Fingerprint Identification Systems (AFIS) with a remarkable success. When tested on large fingerprint databases provided by the FBI and the British Home Office, it delivered > 99% identification accuracy (i.e., < 1% Type I error) and close to zero false alarm rate. In benchmarks conducted by the customer, it achieved better than one order-of-magnitude reduction in the rate of false alarms in comparison with other competing state-of-the-art approaches, especially when presented with partial information, e.g., poor quality latent fingerprints collected from crime scenes or similar uncontrolled environments. ARGTEC proposes to develop and implement a generalized content-based indexing and retrieval system for global information databases. We will expand the ARG technology used successfully in AFIS to a hierarchical ARG scheme capable of processing more generalized global information that contains textual and non-alphanumeric information at multiple levels of detail and resolution. We will show that the hierarchical ARG representation scheme, coupled with our suites of matching algorithms are especially effective in handling noisy and partial information. ARGTEC is working with several industrial partners in both defense and commercial arenas. Our commercialization strategy lies in converting our R&D effort into system innovations, then teaming with large system integrators to infuse it into fielded, operational systems.