{"id":408,"date":"2020-06-04T21:37:58","date_gmt":"2020-06-04T21:37:58","guid":{"rendered":"http:\/\/ssrc.ie\/blog\/?p=408"},"modified":"2020-06-04T21:45:30","modified_gmt":"2020-06-04T21:45:30","slug":"understanding-the-basics-of-facial-recognition-technology","status":"publish","type":"post","link":"http:\/\/ssrc.ie\/blog\/2020\/06\/04\/understanding-the-basics-of-facial-recognition-technology\/","title":{"rendered":"Understanding the Basics of Facial Recognition Technology"},"content":{"rendered":"\n

Biometric security was first commercially introduced to the digital world with the use of fingerprint technology which was to a large extent, grouped under digital forensics. This technology has thrived but can definitely be improved in many ways. Facial recognition technology takes care of the majority of its drawbacks with its ability for remote verification, ease of use, and affordability.<\/p>\n\n\n\n

\"\"
Photo by Ryan Miguel Capili<\/a><\/strong> from Pexels<\/a><\/strong><\/figcaption><\/figure>\n\n\n\n

Working Principles of Facial Recognition Technology<\/strong><\/p>\n\n\n\n

The facial recognition system can be implemented in two major ways. The basic form is 2D imaging which uses flat images from pictures. This is rather inaccurate as the software can easily be fooled by properties such as angles and shadows. 3D imaging, a more advanced system, uses a web of infrared lights that mark every facial characteristic.<\/p>\n\n\n\n

This web of lights is then sent through the camera to the software. The pixels will then be converted into a numeric code. This code is then compared with other stored values in the database in order to make a match. On a more advanced and complex level, some systems use the surface texture analysis <\/a>technique which focuses on skin properties such as wrinkles and distance between skin pores to accurately make a match.<\/p>\n\n\n\n

As with most applications that deal with Artificial Intelligence and Machine Learning, the more the raw data used to train the algorithms, the more accurate the result. Hence, virtually every facial image picked up by facial recognition software ends up getting saved to a database.<\/p>\n\n\n\n

From the technical point of view, facial recognition works based on the geometry of the facial attributes such as:<\/p>\n\n\n\n