Image based Eye Gaze Tracking and its Applications

Anjith George

DOI: https://doi.org/10.48550/arXiv.1907.04325

2019-07-09

Abstract:Eye movements play a vital role in perceiving the world. Eye gaze can give a direct indication of the users point of attention, which can be useful in improving human-computer interaction. Gaze estimation in a non-intrusive manner can make human-computer interaction more natural. Eye tracking can be used for several applications such as fatigue detection, biometric authentication, disease diagnosis, activity recognition, alertness level estimation, gaze-contingent display, human-computer interaction, etc. Even though eye-tracking technology has been around for many decades, it has not found much use in consumer applications. The main reasons are the high cost of eye tracking hardware and lack of consumer level applications. In this work, we attempt to address these two issues. In the first part of this work, image-based algorithms are developed for gaze tracking which includes a new two-stage iris center localization algorithm. We have developed a new algorithm which works in challenging conditions such as motion blur, glint, and varying illumination levels. A person independent gaze direction classification framework using a convolutional neural network is also developed which eliminates the requirement of user-specific calibration. In the second part of this work, we have developed two applications which can benefit from eye tracking data. A new framework for biometric identification based on eye movement parameters is developed. A framework for activity recognition, using gaze data from a head-mounted eye tracker is also developed. The information from gaze data, ego-motion, and visual features are integrated to classify the activities.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

Based on the provided paper abstract, the main problems that the paper attempts to solve are as follows:

1. **Cost Reduction**: Although the existing eye - tracking technology has been around for many years, its high cost has limited its application at the consumer level. The paper proposes to develop an eye - tracking algorithm based on ordinary webcams, without the need for special hardware, thereby reducing the system cost and making eye - tracking technology more popular.

2. **Accuracy Improvement**: The existing eye - tracking technology has insufficient accuracy under real - world conditions, especially in outdoor environments. The paper proposes to develop a new algorithm that can work under motion blur, reflection, and different lighting conditions to improve the robustness and accuracy of the eye - tracking system.

3. **Reduction of User - Specific Calibration Requirements**: Most of the existing eye - tracking algorithms require specific calibration for each user, which limits their wide use. The paper proposes a face - independent eye - movement direction classification framework without user - specific calibration, using convolutional neural networks (CNN) for real - time eye - movement direction classification.

4. **Application Development**: The paper also develops two applications based on eye - movement data:
- **Biometric Authentication**: Use eye - movement parameters for biometric identification, combined with traditional iris recognition technology, to improve the anti - counterfeiting ability and liveness detection ability of biometric modalities.
- **Activity Recognition**: Use the data from head - mounted eye - trackers, combined with self - motion and visual features, to classify indoor activities and improve the accuracy of activity recognition.

Through these methods, the paper aims to overcome the limitations of the existing eye - tracking technology and make it more practical and popular in practical applications.