At the heart of the product line is a hardware component (CM1K) that implements natively the fastest KNN (“K Nearest Neighbor”) implementation known as well as RCE/RBF nonlinear classifiers. This device is not coded, but trained through an on chip automatic model generator for comparing & recognizing 1 vector against N trained vectors. Learning can be real time or off-line. Modules are available to simplify incorporation into your product line.

Both component and image sensor evaluation modules are available to speed your development efforts. Included with the image sensor module (V1KU) is a S/W development tool that adds knowledge building to enable solving your application.

Recognizing your need to integrate the technology into FPGA or SOC devices for certain applications, the CogniMem^™ technology can be made available for licensing. It is a fully digital, CMOS static logic compatible and scalable design.