JeVois makes it easy to add machine vision to your projects by embedding image capture, vision
processing, and delivery of results into the tiny camera itself. Inside JeVois is a powerful quad-core
1.34GHz processor that provides enough processing power to run sophisticated machine vision algorithms
for object detection and tracking, scene recognition, detection and decoding of QR codes or other 2D
barcode markers, face detection, and object recognition using deep neural networks.
Which machine vision algorithm runs can be selected by users or by embedded processors (e.g.,
Arduino) connected directly to JeVois. JeVois can output video streams over USB to send its vision
processing results to a host computer, and simple text strings over serial port, to inform an Arduino
of what it is seeing and recognizing. With a rapidly growing collection of already 25+ machine vision
modules, no programming experience is necessary to start using JeVois immediately -- just download a
vision module from jevois.org, install it on your JeVois smart camera, and enjoy!
Yet, because JeVois software is open-source, you can also create new machine vision modules to run
directly on the embedded processor of JeVois. Development is facilitated by allowing you to compile
JeVois software for any Linux computer, making it easy and fast to debug your new vision algorithms,
later compiling the same code for the embedded JeVois processor.