GazeVisual-Lib

GazeVisual-Lib provides algorithms and datasets to quantitatively evaluate the accuracy and data quality of eye-tracking systems across variable operating conditions such as user distance, head pose, and eye tracker platform movements.

Key Features:

• Algorithms for Performance Analysis: A suite of algorithms to assess accuracy and reliability of eye-tracking systems under diverse operating conditions.
• Visualization Tools: Software to visualize gaze data and analysis results for identification of patterns and anomalies in eye-tracking outputs.
• Benchmarking Dataset: A labeled eye gaze dataset collected from multiple user platforms and operating conditions for comparative evaluation.

Scientific Applications:

• Comparative Performance Studies: Quantitative comparison of accuracy and robustness across different eye trackers and configurations.
• Impact Analysis: Assessment of how user distance, head pose, and tracker platform movements affect eye-tracker performance metrics.
• Data Quality Assessment: Evaluation of gaze-data quality from real-world scenarios to inform downstream analyses.

Methodology:

Integration of algorithms for performance analysis with visualization tools and benchmarking against a labeled eye gaze dataset.