JeVois - A Quad-Core Machine Vision Camera for Makers

Laurent Itti has been aiming forever to build machines that have vision. Itti also expressed unhappiness on part of STEM education that teaches robotics and programming but keep away machine vision completely. Once, he was sure of the fact that no one taught on this topic, he decided to make one of his own. Currently, he is running a Kick starter campaign for JeVois which is an open source camera for maker projects.

The device consists of a video sensor, a USB video port, a quad core CPU, along with a serial port located in 28 cubic centimeter package. He has instilled JeVois with three operational phases for now. A development mode also communicates results and shows up a demo display. The text-only mode of this device shows off text strings from smart camera without making use of USB output. A Pre-Processing mode sends in video for machines to make use of edge maps near interesting objects in a frame. The campaign video shows Itti saying how he wished the system had high frames per second putting picture resolution aside.

The quad-core ARM-V7 processor is clocked at 1.34 GHZ along with a RAM of 256MB DDR3-1600. It also has a dual-core Mali 400 GPU. It has a 1.3 megapixel camera that is capable of running at a speed of 120 frames per second, The campaign page also talks about Windows, Linux, Mac OS X, and Android video streaming applications, The Linux one is working exceptionally well. The open-source nature of this projects sparks special interest as the present limitations are framed out while expecting the users to pick this up completely and run numerous other platforms with it that are not in existence for now.

JeVois is, undoubtedly, an ambitious project that aims at bringing machine vision in makers and educators world. There are several other applications that have been discussed over the campaign page along with recognition of different places, visual attention, reading digits, detection of QR codes, and so forth.