The Center for Artificial Vision
 
Purchase C4AV Products
Total: $0.00
CM-1K
Neural Network Chip

CM-IR2K
Image Recognition Board

CM-SK1K
Starter Kit

CM-PM1K
Prototyping Module

CM-EB2K
Evaluation Base Board

CM-EMB
Embedded Module

CM-V1KU
Vision Sensor Board



Home

Online Payment Service

ISO 9001:2000 Certified
 

CogniMem Technology Ignites Neural Network Rethink

CogniMem is the unique merger and implementation of two reknown essential concepts: a powerful non-linear classifier and a natively parallel architecture.
With the increasing demand for sensory computing (computer perception) and pattern recognition (data mining) in domains such as the industry, defense, security and condition monitoring (human or machine), CogniMem brings a practical solution without the existing limitations of software neural networks such as their speed performance and programming bottleneck.
CogniMem allows a quantum leap in two large domains:
  • Smart sensors with embedded pattern recognition capabilities, thus notifying or tranmitting
    the data or events of interest only. Possible real-time learning.
  • High-speed data mining and unstructured data base search with the ability to match of one
    data set versus any number in a few microseconds.
CogniMem is the convergence between a natively parallel architecture and a highly non-linear classification technology. Its implementation is silicon was made possible by the advances in semiconductor technology (0.13 micron and under).
Technical summary

Robust Well Known Classifiers K-NN (K Nearest Neighbor)

  • Restricted Coulomb Energy (Radial Basis Function neural network)
  • Built-in real-time learning capabilities
  • Supervised and unsupervised learning
  • Management of uncertainty domains
  • Multiple networks and hopythesis generation

Unmatched Performance Parallel architecture allows a recognition cycle independent from the knowledge size

  • Parallel architecture allows daisy-chaining of the neurons, thus unlimited knowledge size
  • Run hundred times faster than an equivalent software model (equiv. to 27 Gigo ops)
  • Allows huge unstructured data mining configuration

Practical and Affordable Small silicon chip (8 x 8 mm)

  • Consumes around a 0.25 watt under 3.3 volts
  • ASIC costs less than $30 in large volume
  • Low-in count connectivity allowing the design of two-chip devices (Sensor plus Cognimem)
  • IP licensing available
 
Please Click Here For Your Free Center for Artificial Vision Catalog
Please Click Here For Your Free C4AV Catalog
Educational Opportunities & Training Materials

CogniMem Technology

Center for Artificial Vision Product Selection Guide

Pattern Recognition Demo

Digital Camera & Pattern Recognition Seminars



C4AV will be joining
with a local Michigan University in the very
near future to provide course work & research
project support.

Center for Global Sensory Intelligence Workshop
 

Google
 

The Center for Artificial Vision accepts Visa, MC, AEThe Center for Artificial Vision accepts Visa, MC, AEThe Center for Artificial Vision accepts Visa, MC, AE

Copyright © 2008 The Center for Artificial Vision All Rights Reserved.

Privacy Statement - Terms of Use