Anti-tracking labeling in rail transit eye-tracking systems involves techniques to anonymize or obscure gaze data to protect passenger and operator privacy while maintaining operational safety and interface monitoring.

Overview of Eye-Tracking in Rail Transit

Eye-tracking cameras in rail transit are increasingly used to monitor operator attention, fatigue, and situational awareness, similar to applications in automotive and aviation sectors . These systems typically employ infrared cameras or wearable sensors to capture gaze direction, blink rate, pupil dilation, and fixation duration, providing insights into operator focus and interface usability . Such data can enhance safety, optimize control panel layouts, and improve human-machine interaction.

Privacy and Anti-Tracking Considerations

In rail transit, privacy concerns arise because eye-tracking data can reveal sensitive information about operators or passengers, including attention patterns, cognitive state, or even identity . Anti-tracking labeling refers to methods that prevent unauthorized tracking or identification while still allowing the system to function effectively. Key approaches include:

• Data Anonymization: Removing or masking personally identifiable features from gaze data, such as facial landmarks or unique eye movement patterns, before storage or transmission .
• Spatial and Temporal Aggregation: Aggregating gaze points into heatmaps or zones rather than recording precise coordinates, reducing the risk of individual identification .
• Edge Processing: Performing gaze analysis locally on the device rather than transmitting raw data to central servers, minimizing exposure of sensitive information .
• Synthetic Labeling: Using artificial or randomized gaze markers in datasets for training or testing purposes, which prevents real operator tracking while maintaining system calibration .

Implementation in Rail Transit

For rail transit operators, anti-tracking labeling can be integrated into driver monitoring systems (DMS) or control room simulators. For example:

• HUD and Interface Optimization: Gaze heatmaps can be anonymized to show which controls or displays attract attention without linking data to specific operators .
• Fatigue and Attention Monitoring: Real-time alerts can be generated based on aggregated gaze metrics rather than individual tracking, ensuring safety while preserving privacy .
• Regulatory Compliance: Systems can be designed to comply with privacy regulations, such as GDPR in Europe, by implementing anonymization and data minimization strategies .

Future Directions

Advances in AI and deep learning allow for joint modeling of gaze patterns and visual context while supporting privacy-preserving techniques . Transformer-based architectures and semi-supervised learning can enable accurate attention monitoring without storing identifiable gaze trajectories, making anti-tracking labeling more robust and scalable for rail transit applications. In summary, anti-tracking labeling in rail transit eye-tracking systems balances operational safety and interface optimization with privacy protection, using anonymization, aggregation, edge processing, and synthetic labeling to prevent unauthorized tracking while maintaining the utility of gaze data .