Progress Report #1 (11/1)
Goals:
- Be able to accurately track the gaze of middle school children using two cameras.
- Develop software to run the eye tracking system.
- Include a method of calibrating the eye tracking system.
- Design a child-friendly case to enclose the system and keep it relatively hidden from view as to not intimidate the subjects.
Hardware:
Two Google AIY kits have been put together and tested using a connection to a phone as well as a connection to a keyboard, mouse, and monitor. The screen allowed us to view exactly what the AIY camera is viewing and see important data on what the camera is “tracking”.
Software:
The Google AIY kits come prebuilt with a smile detection feature which was successfully tested. In addition, we were able to have the kit identify at least 4 faces at a time using a program called pygaze.
To be done:
We are exploring using convolutional neural nets to develop an effective method of tracking gaze. The current plan is to use infrared light as a metric to analyze and track the eye movements. In order to complete this, we will have to buy 2 infrared lights and a infrared sensor. A future plan is to have one camera move to follow the face, so motors and a focusing lens are currently being sought out.