AI-Powered Analysis of Eye Tracker Data in Basketball Game

Area of Science:

• Sports Science
• Human-Computer Interaction
• Artificial Intelligence

Background:

• Understanding visual attention in sports is crucial for performance analysis.
• Wearable eye-tracking technology offers new avenues for data collection in dynamic environments.
• Current methods for analyzing sports-related cognitive processes are limited.

Purpose of the Study:

• To develop and present a novel system for processing eye-tracking data from basketball coaches and referees during live games.
• To extract features from eye-tracking data to understand visual attention patterns and physiological responses.
• To correlate game events with eye-tracking data to infer cognitive processes and decision-making.

Main Methods:

• Utilized two pre-trained Artificial Intelligence (AI) models for data processing.
• Employed the Pupil Labs Neon Eye Tracker, optimized for video analysis.
• Integrated AI models with eye-tracking data to monitor game events and actions.

Main Results:

• The system successfully processed eye-tracking data, extracting relevant features.
• Demonstrated the ability to correlate visual attention patterns with game events.
• Provided insights into coaches' and referees' focus and potential cognitive load via pupil size changes.

Conclusions:

• The developed AI-based eye-tracking system offers a valuable tool for sports psychology and performance analysis.
• Wearable technology combined with AI can provide real-time insights into cognitive processes in sports.
• This research highlights the potential of light neural networks for in-game sports analytics.