Cost Effective Real-time System for cognitive computing using Personalized Eye Blink Detection from Camera.

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Biomedical Signal Processing

Background:

• Eye blink detection is crucial for understanding mental states.
• Current vision-based methods lack personalization and vary in performance.
• Existing eye trackers and glasses are expensive and not widely accessible.

Purpose of the Study:

• To develop a personalized, vision-based eye blink detection system.
• To create an unobtrusive and ubiquitous system deployable with standard cameras.
• To achieve high accuracy in diverse real-world conditions.

Main Methods:

• Utilized a personalized vision-based approach for eye blink detection.
• Implemented the system using standard webcams or mobile cameras.
• Validated the approach on both lab-collected and open datasets.

Main Results:

• Achieved a high F-score of 0.98 on self-collected data.
• Obtained an F-score of 0.91 on an open dataset.
• Demonstrated robust performance under varying light conditions and with/without spectacles.

Conclusions:

• The proposed personalized system offers an accurate and accessible solution for eye blink detection.
• The ubiquity and unobtrusive nature make it suitable for real-world applications.
• This approach outperforms current state-of-the-art methods in terms of accuracy and accessibility.