Commercial Security

Imaging Analysis is arguably one of the fastest growing sectors of the Surveillance market place and rightly is securing its place within Commercial Security considerations. For Shopping Centers, Public Buildings, University campuses and even Commercial premises there is an increasing requirement to secure common areas to a reasonable level. As we see in the British example there is an overwhelming desire to install a greater number of CCTV cameras but who monitors each of these cameras and how is one security guard enabled to identify the threat; a stalker, a murderer, a thief? This is where imaging analysis can assist in a series of measures to ensure that the exception is picked up immediately leading to a chain of events that will ultimately answer the question posed.

Face Recognition Technology

One element of imaging analysis that is rapidly improving is that of facial recognition. The greater the ability to match a series of points on a face related to the dimension and pixel count appropriate between the eyes will result in an increasing percentage success rate in facial identification.

Face Recognition

Humans mostly use faces to recognize individuals. Recent advances in computing technology and capability now enables similar recognition automatically. Early face recognition algorithms used simple geometric models, but the recognition process has now matured into a science of sophisticated mathematical representations and matching processes. Major developments have propelled face recognition technology into the spotlight. Face recognition can be used for both verification (1:1), and Identification (1:N) applications.

Face Template

The heart of the facial recognition system is the Local Feature Analysis (LFA) algorithm. This is the mathematical technique the system uses to encode faces. The system maps the face and creates a "template" as a unique numerical id for that face. Once the system has stored a "template" it can compare it to the thousands or millions of "templates" stored in a database; each template occupies only 2-3 KB of data.

Algorithm

The right algorithm will enable fast and accurate face localization for reliable detection of multiple faces in still images as well as in live video streams. This simultaneous multiple face processing and identification can happen in a single frame. The face needs to be detected in less than 0.1 seconds and then processed in less than 0.2 seconds. This rapid detection rate allows comparison of up to 100,000 faces per second, through the recorded frames generating the collection of the generalized face features from several images of the same subject and writing it to the database. In this way the enrolled feature templates are more reliable and the face recognition quality increases considerably over time.

Identification vs. Verification

Verification (1:1, one-to-one) – This is the process of determining a person's identity by performing matches against one biometric template that is located upon a known ID. 1:1 verification usually uses a manual interface like a card, RFID, code, or any other similar key based indexing.
Identification (1:N, one-to-many) – This is the process of determining a person's identity by performing matches against multiple biometric templates. The identification system is designed to determine identities based solely on biometric information. There are two types of identification systems: positive identification and negative identification. Advanced applications allow our customer full control over recognition method where we support both 1:1 (verification) an 1:N (identification) methods.

Product Solutions

We offers a range of Facial Recognition Solutions designed to provide verification (1:1) and identification (1:N). These start with the basic Logon system that monitors who is using your computer or laptop. Supervisor provides an online check across a classroom of each Logon face and its verified user per computer. Access II and III provide access control support for verification and then combine with Vistrack as a Visitor Tracking system. Taking this to a more commercial level we provide identification solutions integrating Logon, Supervisor, Access and Vistrack into our Suspect and Observer packages. FRS Suspect and FRS Observer are advanced surveillance systems that can automatically detect faces both online or offline using the above fast acquisition methods. They analyse the input from numerous CCTV cameras, comparing the extracted face templates from the cameras with a database containing previously stored profiles and face templates. As each image is detected they automatically raise alarms through whatever interface is required including SMS, PDA, Email and Web.

Conclusion

Face Recognition Technology should be forming a part of your Commercial Security Solution. Used correctly according to position and lighting, a correct and clear image from a camera with sufficient resolution, and integration to other Biometric tools, Access, Time and Attendance, Management Information Systems and of course Imaging systems will provide a comprehensive series of measures to improve your commercial situation. Facial Technology can provide high accuracy levels and with the right pixel count of at least 60 pixels between the eyes enable a 99% read rate.