BANGALORE: Everyday an enormous amount of photos is being uploaded on the social networking sites like Facebook. These photos largely comprise of Facebook users who either upload the snaps or are put it up from another account, reports Demanjo.

Nonetheless, a program by Facebook’s called the DeepFace AI is designed with a facial recognition software that links or matches photos to its users with a 97.25 percent accuracy rating.

However, the program has its set of drawback; the software perfunctorily matches the photos to the users even if it is intentionally not tagged. Tech engineers are utilizing the database library of Labeled Faces in the Wild (LFW) library as a point of reference for their facial recognition software.

Moreover, the new face detection software seems to be a growing need as numerous companies are exploring similar technologies, which includes the government and the notable company Google. This may seems unnecessary and daunting but equal distribution of powers seems quite justified, then giving one group the full authority to exercise.

The in-depth science applied in DeepFace utilizes a technique called Deep Learning that works on a continuous process of examining features in a variety of conditions. It then registers the existing data which analyses and recognizes the patterns in new faces. Thereby it only gets smarter as the DeepLearning program designs a 3D model of the users face. With increasing number of pictures of a user uploaded on the site the more perfect the model gets. Once an accurate model of the user is created, the program can detect or match all the pictures of user in almost all situation or facial expression.

READ MORE: World's First 3D-Printed Jet Engine Developed In Aus

Facebook Updates Feature For Suicide Prevention