Facial recognition technology is developing at an alarming rate, and glass is actually a representative of modern systems and is at the core point of this process.
A recent paper published by the University of Wisconsin-Madison highlights the progress in this field and their “intelligence” Glass can be recognized without sensors or power.” We are using an optical system to compress the normal settings of cameras, sensors and deep neural networks into a thin piece of glass,” the researchers explained. This progress is important because today’s AI consumes a lot of computing power, every time it consumes a large of battery power when you use facial recognition to unlock your phone. The team believes the new glass promises to recognize faces without any power.
Proof-of-concept work involves designing glass that recognizes handwritten numbers.
The system works by light emitted from images of some numbers and then focuses on one of the nine points on the other side that correspond to each number.
The system is able to monitor in real time when the numbers change, for example when 3 changes to 8.
“The fact that we were able to get this complex behaviour in such a simple structure makes real sense,” the team explains.
Arguably, this is still a very long way from occupying any type of market application, but the team is still optimistic that they stumbled upon a way to allow passive computing capabilities built directly into the material, rendering single pieces of glass that can be used hundreds and thousands of times. The momentary nature of the technology offers many possible potential cases, although it still requires a lot of training to enable materials to be quickly identified, and this training is not that fast.
However, they are working hard to improve things and eventually want to use them in areas such as facial recognition. “The real power of this technology is the ability to deal with more complex classification tasks immediately without any energy consumption,” they explain. “These tasks are the key point to create artificial intelligence: teaching driverless cars to identify traffic signals, implementing voice control in consumer devices, and many other examples.”
Time will tell if they have achieved their ambitious goals, but with facial recognition, it is certainly a journey for concerning.
Post time: Oct-09-2019
