Artificial Vision

with NeuroMem neuromorphic chips

Our CogniSight® image recognition technology is an essential enabler for practical articifial vision.

Powered by a trainable NeuroMem® network, CogniSight can learn from examples taught with or without supervision, and describe the recognized objects or events in a semantic or graphical format. A CogniSight engine can be composed of single or multiple networks and take the form of a single low-power SOC, a module or a highly scalable vision system.

CogniSight has applications in miniature vision sensors, industrial and consumer cameras, stereoscopic and hyperspectral systems, DVRs, etc. It can be trained for part identification, defect detection, object classification, target tracking, template matching and more.

Artificial vision requires fast low-power self-learning artificial intelligence. This is why General Vision uses exclusively the NeuroMem® technology for pattern learning and recognition. For more information visit NeuroMem Technologies which promotes and disseminates NeuroMem® accross all markets from edge intelligence to big data.


New demos from nepes on the KocoaFab youtube site

including face recognition, gesture control, and more

New  “Learning Path” information

Easy dive starting with new demos and value propositions, getting down to teaching utilities and chip drivers.

NM500, has arrived! Check the smallest neuromorphic chip, conveniently packaged in a wafer scale package, and delivering 120,000 pattern recognitions per second for less than 125 mW.

New evaluation versions of the Knowledge Builder software suite


Robust AI chip, field-proven and trainable by fishermen, one of the many applications.