iGuide

Inspiration

Motion-disabled people have a hard time working on computers which are widely used these days. We are making a system in which motion-disabled people can use their eyes to select the option on the computer and can use it just like a normal person would use a computer.

What it does

1. Vision Quiz

When a question with four options is displayed, the user can look at the correct answer for 2 seconds to answer the question. The user will then be asked to confirm the answer and if he will look at yes for 2 seconds, the answer would be submitted.

2. Image Analysis for Advertisements

An advertisement image will be displayed for 3 seconds. Later, the heat map containing the analysis of where the user looks the most is displayed.

How we built it

We developed a three-layered architecture to interface between eyetribe device and web application. There are three modules which communicate with each other to acheive the proper synchronization and response time.

Technologies and Hardwares

Eye Tracker Device (Eye-Tribe) : Provides real-time data of user eye-movement on the screen
Socket.io : Excellent technology to handle passage of real-time streaming data (30fps) between multiple levels of clients and servers
NodeJS : Event-driven, Asynchronous backend framework for
Ngrok : Tunneling service to communicate with remote servers
Express : MVC in NodeJS
Mongo : NoSQL database
Web Frontend (Bootstrap) : Responsive grid based frontend framework
Git : Code Versioning
Python : Speedup development process by automating tasks

Challenges we ran into

Real-time streaming data : dynamic, configurable
Multiple levels of clients and servers
Cross-platform compatibilities
Programming language barriers
Sockets and TCP connections: Express handling became difficult
Synchronizing eye blink and other variations
The volume of data (30fps) : Enabling browser support for such high volume data
Sleep : Since we are all humans :p

Accomplishments that we're proud of

Developing an application to make the computer accessible to specially-abled people.
Implementing a general architecture (three-layered server-client/server-client) to make the development of future applications on this platform easier and faster

What we learned

Asynchronous programming to handle real-time high volume data. Managing parallel multiple channels of development

What's next for iGuide

Embedding the eye tracker in the computer itself (no separate hardware)
Creating an accessible API to be integrated with any software
Making control more granular (higher definition screen supports and dynamic coordinates calculation)