NeuroEdge

Trainable pattern recognition on Edge AI Platform

The NeuroEdge is a Raspberry Pi based AI development platform that uses a neuromorphic device.

It supports a variety of sensors that can be used for AI systems.
It also enables real-time field training and inferring on devices that don’t require a network environment.

What is inside of NeuroEdge?

NeuroPi

Block Diagram

SPECIFICATION

Components

Neuromorphic	2 * NM500 (up to 1,152 Neuron)
Weight	190 g
Dimensions	95.00 * 68.00 * 55.00 mm
Bracket	VESA 75 * 75 mm, 100 * 100 mm

Hardware Specification

SoC	Broadcom BCM2711 SoC
CPU	ARM Cortex-A72 1.5GHz
GPU	VideoCore VI 500MHz
Memory	2 GB LPDDR4
Storage	32 GB (Micro SD)
Neuromorphic	2 * NM500 (up to 1,152 Neuron)
Network	Ethernet, Wi-Fi, Bluetooth
Video output	Micro HDMI * 2
Audio	3.5 mm Jack
USB	USB 2.0 * 2 / USB 3.0 * 2
GPIO	40 Pin GPIO Expansion Header
Power	5 Vdc (more than 2 A)
Operating System	Raspbian Buster

Manual/Software

NeuroPi H/W USER GUIDE

GETTING STARTED WITH NeuroEdge

NeuroPi SDK DOWNLOAD

Application Note

Face Recognition

Simple NM Detector

NeuroEdge Installation with Bracket

Bracket of NeuroEdge is compatible with the VESA MIS-D standard.
Computer monitors and small televisions usually have the VESA MIS-D standard. With the VESA MIS-D standard, the mounting holes on the back of the monitor or TV have a horizontal and vertical distance of 75 or 100 mm to each other, which corresponds to a VESA standard of 75×75 or 100×100.
NeuroEdge can be connected to the VESA MIS-D standard.