FAIL (the browser should render some flash content, not this).




FAQ | Face Recognition
Print Email
Where can iQ-Face be implemented?
iQ-Face can be implemented in various security, marketing and customer service applications. For example IQ-Face can be implemented to recognize known shoplifters as they enter the store or VIP clients of a hotel even before they reach the reception desk. It can also be used to find matches from a pre-existing database for images of faces recorded at a terrorist incident.

Is this Technology unique to iOmniscient? 
The core technology for Facial Recognition in a Crowded Scene is internationally patented. Further it is surrounded and supported by other innovative technologies which are unique to iOmniscient. It is recognized as the first (and at this stage the only) effective "many to many" Face Recognition system available.

Is there any limit in terms of the numbers of faces to be enrolled?
Theoretically there is no limit.

How fast can faces be recognized?
This would depend on the size of the database. With a database of over a few thousand faces the system would perform facial recognition in less than one second.

What is the accuracy of IQ Face?
iOmniscient's system is designed for many to many facial detection and matching to a images in database for eventual recognition and/or identification. It is able to provide 80% accuracy which is significantly higher than what any human can achieve in a similar environment and much more than any comparable system in the market.

Are the recognition algorithms an extraction of the facial characteristics?
Or does it include other algorithms?The system understands many different characteristics of the face such as the bone structure, the position of the eyes and others.

What must the angle of the face be for IQ to recognize?
The system will work as long as both eyes can be seen. In most situations this means that the face can be tilted in any direction for up to 30 degrees and the detection would still be reasonably accurate.

What is the pixel numbers needed for face detection?
The system requires 22 pixels between the eyes for accurate facial detection and recognition. Images of up to 8 pixels can be processed; however accuracy of facial identification would be affected.

Is the system limited by racial characteristics such as skin colour?
No, the system is not limited by these characteristics.

What is the maximum velocity of people walking or in vehicles to be detected and successfully identified by IQ-Face?
These variables would be dependent on the camera used in the application. IQ-FACE requires a picture with a clear resolution with at least 22 pixels between the eyes for effective identification of faces. Therefore camera features such as shutter speed and mega-pixel capabilities would be important to enable clearer pictures to be captured allowing for higher velocity facial detection and recognition.

What differentiates this system from other Facial Recognition Systems?
Other systems require the subject to be in a controlled environment with steady lighting. Further they require 300 pixels between the eyes and sometimes require more than one camera view. This makes them computing heavy and unusable in practical environments where the technology is required to be used with standard security cameras.