Luxand, a software development company with core expertise in artificial intelligence and biometric identification, has released its advanced face recognition engine, the Luxand Face SDK 1.7.

 

The new library can be used to design biometric applications, security and access control systems and includes a unique facial recognition tool that can detect facial features of each face and can compare faces feature by feature.

 

Luxand is focused on the development of solutions and SDKs for automatic recognition of human faces and facial features. The company claims that their SDK is not like the technique of digital cameras that run on Digital Signal Processors (DSP) used in face detection algorithms. Though these cameras have better auto focus and proper exposure, the company says they can only recognize the face and not facial features- making their application especially important.

 

Luxand’s facial feature recognition solution automates common post-processing tasks such as red-eye removal and skin tone adjustment and allows users of digital cameras to benefit from more precise automatic adjustments and touch-ups that don't require user intervention.

 

Luxand FaceSDK employs sophisticated algorithms to detect frontal human faces on digital pictures quickly and reliably.

 

The tool employs two modes in facial identification. In the first mode (1:1 – one to one), the algorithm verifies that the person is who he claims to be by matching a person's face to one face template in the database. The second mode (1: N – one to many) helps to identify who the person is by comparing the face against all face templates stored in the database. Then the SDK processes the image, detects human faces, and returns the coordinates of 40 facial feature points including the eyes, eye corners, eyebrows, mouth corners, nose tip, and so on.

 

The identification accuracy of the Luxand's algorithm has been evaluated in FERET tests and found to be the highest in the category as compared to the algorithms implemented in competing commercial systems.

 

"The results we've seen prove our product to be the best in class," said Alex Konoplev, CEO of Luxand. "Our technology can unlock the potential for developers to make software products with facial feature recognition, which is truly outstanding in terms of accuracy and speed. Luxand Face SDK dramatically reduces the time and costs of creating such an application and can be easily integrated into the software project."

 

This particular Face SDK can be used to allow access to laptops or PCs using ones face or to assist law enforcements in identifying a suspect. For example, in the airport or in the street, the security officer can capture a face of the detained person and check the photograph against a database of previously documented criminals and suspected terrorists. In a matter of seconds, the systems will tell if the person in question is among those on a watchlist for terrorism or crimes.