Traditional Algorithms

• Early facial recognition systems looked to create "matches" between a source image on file and the image being captured on the scanner. Different matching algorithms have been developed over the years to identify specific features and angles on the face and the proportions and relationships between them. Some algorithmic methods include the Generalized Matching Face Detection Method, Perturbation Space Method and Adaptive Regional Blend Matching.
  
  

3-D

• The newest facial recognition technologies involve 3-D scans of the users' face. This measures the specific contours of the eye sockets, or nose, and eliminates disruptions that could occur from changes in lighting, expression or angle.
  
  

Crowd Identification

• Advances in facial recognition software now allow for security experts to identify faces in a crowd of people, without their consent and without their knowledge. This is a clear advantage over many other forms of biometric authentication like fingerprinting and iris scanning, which require the willful participation of the user.