Audio & Video Indexing Services (StreamSage)

GMSI serves as the federal systems integrator and professional services provider for StreamSage Inc.’s indexing and language translation solutions. StreamSage products listen to and watch audio and video content to develop a contextual understanding of the content. GMSI affords its customers access to the StreamSage search engine to find relevant content and display it in a prioritized list of search results, saving time and money by rapidly extracting needed information. The resulting level of automation and granularity provides enterprises with unprecedented management and control of access to the information contained within their audio and video content. The StreamSage product is especially applicable to government, security, law enforcement, corporate, broadcast media, legal, scientific, and academic environments.

Audio & Video Capabilities
GMSI offers several audio and video (A/V) capabilities that afford user advantages:

• Enable granular search of archived A/V files through the automatic editing of content relevant to the interests of the searcher.
• Eliminates today’s expensive process of first manually tagging A/V content and then editing that into useful information (e.g. video clips) to create a searchable index of audiovisual information.
• Automates media ingestion from distributed locations.
• Alerts analysts of significant mentions of security issues in surveillance data, intelligence briefings, or any other type of A/V content.
• Personalizes delivery of intelligent A/V content for the analyst, war fighter, or other personnel, along with any corresponding matched text documentation.

Core Technologies
Notably, two of the core technologies being applied by StreamSage were listed in MIT Enterprise Technology Review (February 2004) among 10 emerging technologies viewed as changing our world in the near term. It is singularly impressive that these two StreamSage core technology applications, universal translation and Bayesian machine learning, are in the company of leading edge advances like turning cells on and off via synthetic biology,
scanning materials for flaws with T-Rays, and “curing” rare
genetic diseases through personal genomics. GMSI affords its customers access to StreamSage core technologies:

• Automated “Relevance Interval” Indexing of Audio & Video Content
  StreamSage technology employs path breaking computational linguistic and statistical analyses (developed under projects funded by NIST, NSF, the Air Force, and commercial customers) to automatically identify precise, relevant, clips of content within audio and video sources (e.g. surveillance, briefings, news media, conference calls, etc.) and to correlate relevant A/V with relevant text. Developed with recognition that people are limited in their ability to organize and integrate data from multiple sources, StreamSage applies Bayesian machine learning techniques in enabling content indexing solutions to be drawn from complex webs of interconnected dependent video and audio data. No other media analysis system provides this automated editing and manipulation of rich media, which has been commercially proven to dramatically improve the efficiency with which information can be found and reused within audio and video content.
• Unambiguous, Multilingual Information Extraction
  Combining components from StreamSage and PARC (Palo Alto Research Center), the multilingual information extraction system has an unprecedented ability to identify facts, relationships, and events from text, speech, and structured data that are relevant to a given user or decision. This allows for the automated population of decision support templates with assertions that are linked back to source documents. StreamSage researchers are in the forefront of escalating efforts to utilize mathematical models and natural language processing methodologies to increase the accuracy and efficiency of computerized translation. StreamSage is implementing advances in semantic analysis to extract the most likely meaning of any text or speech and express it in another language. These technologies are opening the door to universal translation for international business meetings, surveillance, intelligence gathering, and document research.
• Domain Scalability
  Because the StreamSage language modeling, machine learning, and sense disambiguation systems are fully automated and eliminate much of the ambiguity that requires domain-specific manual training, systems can be adapted to new types of scenarios, decisions, information sources, and threats. This is particularly important, as it addresses one of the largest barriers to applying previous natural language processing technologies.

Contact us today to learn more about StreamSage and how GMSI can apply our experience and expertise to cost-effectively meet your audio/video content management needs.