Home > News & Events > MITRE Publications > The Edge >

When my orders came in to go to Israel at 11:00, my pulse raced. I was being sent at the last minute to support important trade negotiations with Israel because our Israel trade expert was away on another mission. As the Middle Eastern trade analyst, I was next in line.

With only a few hours before I had to leave, I needed a quick way to get a sense of recent public opinion on Netanyahu's trade position. If public opinion was not favorable, Netanyahu would be in a much weaker bargaining position. What were Israeli news and CNN reporting? I started Broadcast News Navigator (BNN) and asked for all stories about Netanyahu for the last three weeks. I was presented with the screen shown in Figure 1.

BNN found twenty-two stories involving Netanyahu. The story summaries helped me quickly find the stories that had the information I needed. I examined the closed caption transcript for one story containing an extended clip from a Netanyahu press conference where he was convincing the Israeli senate that his hard-line stance was correct. His answers to pointed questions should prove very useful!

While reading one of the stories that Broadcast News Navigator returned, I noticed the mention of heavy fighting between the Israelis and Palestinians. Realizing I may be going into a dangerous location, I requested from BNN a trend report showing the most common subjects of news stories over the last three months. From the report, it appeared that fighting had steadily increased the last few months. One story talked about fighting in the region where the trade talks were going to be held. From the story, I noted the cities in that region I should avoid.

In the above scenario, the analyst showed one way Broadcast News Navigator (BNN), a news exploitation tool created by MITRE, could be used. BNN is a web-based search tool that allows an analyst to query and view a large collection of broadcast news. BNN's unique approach of capturing, annotating, segmenting, and summarizing the news into individual stories allows an analyst to drill down quickly to the desired news information. BNN is currently in production and automatically processes new broadcast news stories every evening.

Besides the powerful time-based story and trend analysis search capabilities discussed in the scenario, BNN provides three other robust search interfaces: search by tags, a free-text keyword search, and a topic-based search. First, the search by tags capability allows an analyst to search for stories containing the automatically detected person, location, or organization tags in a story. Second, the free-text keyword search returns stories containing the chosen keywords. Finally, the user can search the database by selecting a story topic (such as "Chinese Leaders," "Israeli women," "Swiss Banks") and viewing the related stories.

The BNN tool resulted from MITRE's work in collaboration with research labs of companies and universities. The creation of the BNN tool is funded by the Community Open Source Program Office (COSPO). This tool is currently being installed on the COSPO OSIS (Open Source Information System) network (see March issue of The Edge for an explanation of OSIS).

BNN interacts with multimedia data gathered from the Broadcast News Editor (BNE) system. BNE takes video input and processes the imagery, audio, and closed-caption channels of information using annotation techniques built for the particular channel. With the channels of information and annotations, BNE correlates the data to detect broadcast, commercial, and story segments. With the segments, BNE generates a gist, theme, and key frame for each story and stores the information into a Multimedia Database Management System consisting of Oracle's Relational Database Management and Video Server System. Through Oracle's web server, BNN is able to query the Multimedia Database Management System.

---

For more information, please contact Andrew Merlino using the employee directory.

---