Summary
The Australian company, BrainChip Holdings Ltd., has developed an Autonomous Visual Feature Extraction system (AVFE) named SNAP based on a technology named spiking neural processor technology. It can process 100 million visual events per second, without any initial knowledge about the content, the processor learns autonomously and starts to find patterns in a set of streaming images. This can be used for applications such as collision avoidance in autonomous vehicles and drones, surveillance and medical imaging.
SNAP currently uses an artificial retina named “Davis Dynamic Vision Sensor” to obtain streaming video information. Instead of outputting frames of video, Davis outputs signal spikes whenever the contrast changes in a pixel. Suitably modified radar or ultrasound streams can also be used as information streams. SNAP does not require human supervision to function. BrainChip Holdings Ltd is expecting to commercialize this processor n near future.
Increasing popularity in development of custom designed processors to achieve specific functions
Implementation of algorithms in hardware level increases efficiency.
Increasing trend of intelligent machines that do not require human interbention.
SNAP is a neural processor technology that can recognize patterns in a stream of images autonomously.