Eye/head tracking technology to improve HCI with iPad applications

Area of Science:

• Human-Computer Interaction (HCI)
• Assistive Technology
• Computer Vision

Background:

• Existing human-computer interaction (HCI) methods for individuals with special needs can be limiting.
• The need for intuitive and accessible control methods for mobile devices like the iPad is growing.
• Advancements in computer vision offer new possibilities for non-traditional interaction techniques.

Purpose of the Study:

• To present an alternative human-computer interaction (HCI) method for individuals with special needs using an iPad.
• To develop and validate an eye and head tracking system leveraging the iPad's front camera.
• To assess the system's accuracy and usability across diverse user groups.

Main Methods:

• Utilized the iPad's front camera for real-time eye and head tracking.
• Employed OpenCV and the Haar Cascade algorithm for facial feature detection.
• Conducted tests with 22 users of varying ages and characteristics to evaluate system performance.

Main Results:

• The system achieved an overall accuracy rate between 60% and 100% across three test exercises.
• Haar Cascade detection was highly effective for faces (100% accuracy), though less so for pupils due to lighting interference.
• Demonstrated that user constraints do not necessarily impede technology enjoyment or usability.

Conclusions:

• The developed eye and head tracking system shows promising results for improving iPad accessibility for users with special needs.
• The Haar Cascade algorithm effectively detects faces, forming a reliable basis for gaze-based interaction.
• Further development and updates are recommended to enhance system capabilities and expand its application.