Content-Based Multimedia Indexing and Retrieval System (MIRS) for Web Database Management

"One of the critical capabilities to enhance the availability and performance of C4I systems for Homeland Defense and other major Air Force initiatives (e.g., DMFE, TAS, DICE, Broadsword, JBI, SSW and BP) is indexing and retrieval of large multimedia,multi-source information. Current indexing and retrieval techniques in these systems are based on limited sets of keywords and simple lookup tables. ARGTEC is pioneering the use of a new, innovative technology for content-based indexing and retrieval inlarge information databases. During Phase I of this program, we proved that our technology is extendable into the domain of multimedia content-based indexing and retrieval. We also developed a proof-of-concept demonstration for content-based indexing andretrieval of free-text documents. Our demonstration delivers very impressive results with very high selectivity and reliability in real-time retrieval of free-text messages. Previously, we have been successful in applying this technology to AutomatedFingerprint Identification Systems (AFIS). In tests on large fingerprint databases, it delivered > 99% identification accuracy (i.e., < 1% Type I error) and close to zero false alarm rate.ARGTEC proposes to develop and implement a content-based multimedia indexing and retrieval system for web database management. We will expand our ARG technology to real-time processing of more generalized global information containing free-text as well asmultime