Related Products
- CME2K - CogniMem module with hirose
interconnect (from 2048 neurons to 12,288
for a stack of 6). - Other resources
Related Documents
- Target Tracking with the ZISC, the ancestor
of CogniMem - Adaptive Optics Video showing V1KU
replacing a PC for pattern matching. - KNN in microseconds a technical paper
describing the performance of CM1K in
KNN mode. - Infrared Target-Flare Discrimination using
ZISC (ancestor of CM1K) a technical paper
courtesy of Royal Military College. - Onboard Feature Indexing from Satellite Lidar
Images, a technical paper referring to ZISC,
the ancestor of CM1K. Courtesy of Carnegie
Mellon University
Tracking objects of interest and detecting an event real time are naturals for the CogniMem technology. Practical distributed detection by many sensors can be made a reality with this technology. Real time event detection saves the user communication bandwidth as well as monitoring resources. Applications are numerous in the defense industry wherein neural network technology has been embraced. CogniMem is a direct hardware implementation.
Read more from an expert
Abstract from the book “Target Tracking with the Zero Instruction Set Computer” released in 2010 by Wendall R. Deck. ZISC is nothing more than the ancestor of our CM1K chip. (“ancestor of our CM1K chip” should link to Company Profile/Background).
“The Zero Instruction Set Computer (ZISC) is an integrated circuit containing a Restricted Coulomb Energy neural network that employs parallelism to execute many operations in very short time periods. We evaluate the computational capability of this device in the context of using it for image pattern recognition, seeking to consume less time than a computer without the ZISC. We discuss the feasibility of using a correlation coefficient calculation to determine if a portion of an image is a match to a given pattern, and describe a serial algorithm to use on a computer without the ZISC. We also discuss four algorithms that are implemented on a computer that accesses a ZISC via the PCI bus. The results of our experiments show that the performance of our ZISC-based algorithms is indeed an improvement over the serially-implemented algorithm.”
Related CogniMem Technologies


