Facial Recognition systems are designed to match a persons face in two or more separate pictures to verify that they are the same person.
What it’s used for
Facial Recognition has been used for a variety of different purposes. One such use involves legal matters, where the technology is used to match still shots from videos of criminals with a picture to confirm their identity. It is also used for fun services on some websites such as MyHeritage in order to compare your face with celebrity faces and tell you who you most look like.
How it works
The most often used Facial Recognition software on the internet relies on 2D images, which are compared. This typically means that you must be looking directly at the camera, and changes in angle of the face and lighting can drastically change the results.
Facial recognition software is designed to first recognize a face and extract it from the rest of an image. It then measures various features of the face, including
- Distance between the eyes
- Width of the nose
- Depth of the eye sockets
- Shape of the cheekbones
- Length of the Jawline
These and other defining points of the face are measured in numerical code, called a faceprint (like a fingerprint) which represents this face in the database. This numerical code is compared to the numerical codes of faces already stored in the software’s database.
A more accurate technology is being used in more professional settings. This software compares two 3D images of a person face, and targets points such as bone structure for a much more accurate comparison. The software detects a persons face in a picture, and then determines the head’s position, size and pose. This software can match a persons face with the correct one in its database in poor lighting, and up to a 90 degree angle away from the camera (a facial profile).
The rapidly advancing technology by police and federal agencies has posed privacy concerns from the general public and federal government. The police will use the system to identify criminals by matching pictures against mugshots recorded in law enforcement databases.
Sean mullin, CEO of BI2 Technologies, the creators of MORIS (Mobile Offender Recognition and Information System), has activiely tested his device to various local, state, and federal law agencies. He said that he wasn’t worried about the federal government deciding this was a privacy issue, since the device is hooked up to law enforcement databases, not public ones.