According to a paper called DeepFace that was published on Facebook’s site, the social network has created facial recognition software that has a 97% success rate.

While many facial recognition programs are far from accurate, and can be tricked by poor lighting, make up, etc, Facebook has had plenty of access to pictures which has led to the creation of this program.

“This deep network involves more than 120 million parameters using several locally connected layers without weight sharing, rather than the standard convolutional layers,” it says in the abstract for the study. “Thus we trained it on the largest facial dataset to-date, an identity labeled dataset of four million facial images belonging to more than 4,000 identities, where each identity has an average of over a thousand samples.”

“The learned representations coupling the accurate model-based alignment with the large facial database generalize remarkably well to faces in unconstrained environments, even with a simple classifier. Our method reaches an accuracy of 97.25% on the Labeled Faces in the Wild (LFW) dataset, reducin,” the abstract continues, “the error of the current state of the art by more than 25%, closely approaching human-level performance.”

Facebook just keeps getting better and better.

[via Privacy SOS, DeepFace Abstract, image via Thos Ballantyne’s flickr]