Attention Makers

CATEGORY: SOFTWARE

Gaze IO System

MAKERS: Asheesh

We are a team of undergraduate software engineers developing a system which aims to enable desktop navigation and control taking the user's eye gaze as an input. The distinction that makes this project unique is that unlike conventional gaze tracking systems, it does not make use of any high resolution cameras, tracking devices or special lighting. It solely relies on the images captured through the onboard webcam. This makes it an ideal application for integration into existing computers with built in cameras, without any additional investment. We believe that as the system evolves, it will make current commercial eye tracking suites, that rely on external hardware, obsolete. We intend to make the system easily deployable for an improved user experience. To accomplish this, the software utilizes various algorithms at different stages of the process, which are being gradually evolved to run efficiently with minimal run times. The software is being developed in with the OpenCV framework, and uses OpenGL, OpenMP, and Gtk. The 2D gaze estimation software will work as a single application emulating an input device that will interact and control the desktop environment by performing basic manipulations, like navigation, and application manipulation. Initially, the system will be targeted towards Linux based distros. Currently, we are continuing development at a brisk pace. We are developing a time efficient algorithm for high accuracy eye detection, and testing functions that estimate the position vector. The project should be fully functional by May 2015. Github Repo - https://github.com/asheeshr/gaze-io-system

Home Previous Next

Vote Share Comment