ANM5000, neuromorphic chip with 5500 neurons
Opening new frontiers for smart sensors and cognitive storage. Can learn and recognize patterns extracted from text, measurement, images and signals with orders of magnitude less energy and complexity than modern microprocessors.
In addition, the chip has digital video input port directly connected to a built-in recognition logic which subsamples a region of interest and broadcasts it after each frame to the neurons for immediate recognition.
The default interface to the chip is a 26-pins bidirectional bus. The chip also features an I2C I/O port for a slower but easier access to the neurons.
|Neuron memory size||256 bytes|
|Category register||15 bits|
|Distance register||16 bits|
|Context register||7 bits|
|Recognition status||Identified, Uncertain or Unknown|
|Classifiers||Radial Basis Function (RBF), K-Nearest Neighbor (KNN)|
|Distance Norms||L1 (Manhattan), Lsup|
|Recognition logic||Subsampling of 1 user defined ROI per frame|