Abstract

Keywords: wearables, anthropometry, open-source

Concept

Methods: Raspberry P'Eye-tracker

A gist on GitHub \citep{gist} details this software process, including the specific lines of code used.

Raspbian

To get started, we installed Raspbian Jessie \citep{Rasbian} on our Raspberry Pi.

Keyboard Layout

The physical keyboard I was using is a US-layout keyboard, but Jessie chose a default layout of GB. I do not know off the top of my head where things like pipes and tildes live on Commonwealth keyboards, so I changed the driver layout to match the physical labels.

OpenCV

From Adrian Rosebrock's 2015 October 26 pyimagesearch guide "How to install OpenCV 3 on Raspbian Jessie" \citep{HowToOpenCV}, we worked on prepping the Raspberry Pi for computer vision. On the first time through, the battery we were using to power the Raspberry Pi ran out before OpenCV was done making. Another computer vision library, SimpleCV \citep{simpleCV}, which is lighter weight than OpenCV, appears to be sufficient for what we are trying to do. In the interest of time, we installed SimpleCV instead of OpenCV.

SimpleCV

For OpenCV, we were installing into a virtual environment; for SimpleCV, we did the same. Following the instructions provided in SimpleCV's official documentation \citep{SimpleCVInstall}, we were able to get SimpleCV running on our Raspberry Pi. Though our SimpleCV was running, we were consistently getting the following error: