NeuroPi
Trainable pattern recognition for Raspberry pi

The NeuroPi is easy to use on Raspberry Pi for evaluating user application on a neuromorphic device called NM500, which has 576 neurons, 256 bytes of memory, and hard-wired logic per neuron to achieve functions mainly required in user application. (NeuroPi has two NM500s) It begins to evaluate user application, such as pattern matching and machine learning using NM500 with the provided software development environment.
Block Diagram

SPECIFICATION
Power | 5V (via GPIO Connector) |
Neuromorphic | 2 * NM500 (up to 1,152 neuron) |
FPGA | LCMXO3LF(Lattice) |
Board Interface | SPI (up to 15 MHz) |
Support Platform | Raspberry Pi |
SDK | C/ C++ , Python |
Size | 65 * 30.75 (mm) |
Weight | 15 (g) |
