IBM’s been busy. A couple weeks back they unveiled a new line of neural computing chips that try to mimic the kinds of connections found in the brain. It could unleash incredible new leaps in parallel computing and revolutionize artificial intelligence. Samsung, it seems, wanted to leverage that new processing power as soon as possible. The electronics giant announced this week that they’ve been working on an image processor that can record in 2,000 frames per second.The processor, which IBM dubs TrueNorth, uses 4,096 separate processing cores to form one standard chip. Each of them can operate independently and are designed for low power consumption. Samsung hopes that the chip will help with visual pattern recognition — something our brains excel at. As such, they’ve incorporated their image sensor and a TrueNorth chip into a gesture recognition system which is already performing spectacularly well.This works because each pixel on the sensor operates independently and only reports changes it sees, much like a compressed video. That means it requires a fraction of the processing power and resources most cameras need. With less overhead, Samsung can push the sensor to record much, much faster — almost ten times that of a standard, 250 fps high-speed camera.Samsung wants to use the tech for safety features for autonomous cars or more advanced gesture recognition. The insanely high speed of the camera could allow it to detect much subtler changes than modern systems. Being able to improve the responsiveness of an automated vehicle with faster systems is also a top priority.It’s just one more example of how we’re already living in the future.