NCJ Number
195391
Journal
Law Enforcement Technology Volume: 29 Issue: 5 Dated: May 2002 Pages: 12-14,16
Date Published
May 2002
Length
4 pages
Annotation
This article describes the procedures of facial-recognition technology and how it can be applied in law enforcement operations.
Abstract
In using facial-recognition technology, a person's picture is taken, with either live video or still cameras. The image is then compared to a database of facial images for a possible match. The computer then yields all the possible matches in about 1 second. For the process to yield maximum results, the picture must be of good quality under adequate lighting. Once an accurate image has been taken, it is converted into a strand of numbers. This code is compared to all the other codes (images) that are stored in the database. The matching process involves looking for special relationships between different "landmarks" in the face, which include bone structure, the distance from the tip of the nose from the face, the bridge of the nose, the depth of the eye sockets, and the structure of the cheek bones. Visionics uses about 80 points on a suspect's face when comparing them to the database. The system requires between 14 and 22 matches for a candidate image to be selected. The features matched in other commercial software are also described in this article. Regarding law enforcement operations in which the facial recognition technology may be used, the most controversial is its use as a surveillance tool. Americans entering airport checkpoints, sporting events, and other gatherings are not usually aware that their pictures are being taken for database matches. Other uses include the booking of a suspect to determine if he/she may be a wanted fugitive, for victim use in identifying suspects, and for limiting access to high-security areas. 5 illustrations