Computer vision and 3D sensing company, Alcatraz AI has announced that it has received USD4 million in funding for its Facial Recognition software and hardware that controls physical access to government buildings and corporate campuses. Based in Palo Alto, California, Alcatraz AI utilizes Machine Learning and computer vision for its Facial Recognition software. The company integrates software with a hardware device equipped with sensors that mounted next to a locked door to identify a person’s identity as they approach the device and open the door in less than one second.
According to the reports, the fund will be utilized to grow Alcatraz’ computer vision, advance underlying algorithms, and reduce its false acceptance rate from 1 in 250,000 today to 1 in 500,000. Alcatraz AI was created by former Apple engineer Vince Gaydarzhiev, who helped build Face ID on iPhone 10. In a statement, Gaydarzhiev stated that Pilot programs are currently underway at companies like Google and SAP, as well as with multiple police departments in the San Francisco Bay Area. As equipped with sensors for color and badge detection in addition to 3D mapping, the company’s device runs on NVIDIA TX2, a graphics processing chip (GPU) often utilized to power edge AI deployment in drones and delivery robots. Additionally, the 3D mapping sensor recognizes the physical presence of a person in front of the door, thwarting people from holding up a photo of an employee to trick the system and then gain access.
Alcatraz device, rather than the need for an enrollment plan to record the faces of each person before getting started, can automatically generate profiles for each employee. For this, the system firstly requires an employee to scan their ID badge, but after a couple of days, it enables the person to unlock doors with their face alone. As part of the development, Gaydarzhiev noted that Alcatraz is also working to make its system talk with APIs to incorporate things like blacklisted databases of criminals or former employees so security can be notified through existing infrastructure management software if someone unwanted tries to get in.