One person, multiple identities
Crime and terrorism attempts are pervasive and increasingly sophisticated. Yet they share one thing in common...they involve people, making identity recognition—knowing who a person is and who they associate with—a critical weapon in fighting crime.
In our global, mobile society, identities are easily "blurred." Organizations merge and combine databases. People get married or divorced. Names vary across cultures. Clerical errors are made once and perpetuated. Opportunity is ripe for people looking to misrepresent themselves for purposes of theft or fraud.
Further complicating the situation is the fact that organizations find themselves limited in their ability to respond due to privacy concerns.
With IBM Entity Analytics, you can balance your "need to know" about an individual, with his or her right to privacy. This data engine can sift through a wide spectrum of databases to build a deep, timeless "view" of a person, including relationships. This view changes in real time as new information is introduced into any one of the databases. Details are accumulated—not just updated—so the profile retains access to old addresses and employers, etc. even after original data has been purged from the source database.
Global Name Recognition
What's in a name?
While numbers can be verified, names can't...making them one of the easiest and most common means of "blurring" an identity. IBM's Global Name Recognition solution, which includes a database with almost one billion names from around the world, can help organizations recognize and understand the person behind the name.
On July 7, 2005, a series of terrorist bombs hit London's public transport system during the morning rush hour, killing 56 people. Elements of London's video surveillance system were critical in helping police to identify the bombers and solve the case. Today, one in four major cities in America are investing in video surveillance, making it the fastest growing segment within the electronic security industry1.
Rewind back to ten years ago Surveillance meant a security guard sitting in the basement watching over a bank of closed-circuit TV screens, changing the tapes every six hours. To stop a crime, he would have to be watching as it happened. To identify the incident would take hours of tracking. And linking that video to related events, such as gunshots fired, would be exhaustive.
Today, with IBM Smart Surveillance, that same guard might be out patrolling while carrying a laptop or PDA, with access to hundreds of high-res camera feeds and alerts that signal events as they happen. He could search through hours of video in moments, and identify patterns of activity over the long term. Other data, from sensors on the same network or extracted from the video, could integrate with license plate readers, facial recognition or gunshot detection systems to fill in the complete story.
1. MSNBC, "Tech: Surveillance cameras become big business", March 15, 2006
Face cataloging: who is where? In this store, the video surveillance system answers the question "who is where?" by using a multi-camera set-up, a Face Cataloger system and 3-D position tracking. Once a person is detected, the cameras automatically pan, tilt and zoom to acquire close-up frontal views of the person's face. The face images are then associated with the 3-D tracking information, which identifies where the person has moved. The face images can be used for identification by a person or a computerized system. In standard surveillance, the face is often missed as a person passes the camera.
Find the red cars, show me the people In this hypothetical example, a bank robbery took place, with perpetrators fleeing in a red car. Investigators can requisition footage from the bank's outside surveillance cameras and query "show me all red cars that passed by the corner at this time." This can narrow down the search to a specific car model. License Plate Recognition, under the right conditions, can apply optical character recognition to extract the plate number. In addition, Smart Surveillance can analyze a secondary query: at what time were there fewer or more people?...with the goal of identifying potential witnesses.
