Even though identical twins are sometimes distinguishable by the human eye, facial recognition algorithms will not be able to pinpoint and distinguish the physical traits between each of them.

Facial recognition looks for the most similar looking face in the database and the results provide a list of candidates. The displayed results very often vary in rank because of a wide variety of reasons. Among them are original photo quality (high-quality image vs. low-quality image), the size of the database you are searching against, and the filters applied to the search. More important to note, facial recognition results still require a human verification from the user.

With twins, the human analysis requires a user to focus on minute differences such as scars, facial moles, blemishes, age spots, and even markings such as tattoos. This human aspect in the facial recognition process is known as facial identification. Facial recognition is not absolute. Because of this, the user must verify that the candidate meets physical similarities and still requires a user to perform a secondary level of verification by performing a background check to validate the possible match candidate. It is important to remember that this process is best practice whenever facial recognition is used for investigative reasons. Once the validation has taken place, the possible match is only an investigative lead. If twins were part of an investigation and one is identified through facial recognition, that would be considered a high-quality lead and enough for law enforcement to investigate further.

Race-based facial recognition is not prevalent in most facial recognition applications today and Vigilant Solutions does not recognize race through our searching algorithms. Facial recognition algorithms identify facial landmarks known as nodal points. Among them are the eyes, nose, corners of the mouth, and chin.