Gaze Tracking Acceleration using CUDA Technology

Low-cost gaze tracking systems are in great demand due to their wide range of application. Commonly, extra devices are needed (for instance, head mounted cameras); however, in this investigation gaze tracking is performed in real-time based on the video stream from an infrared video camera. A comparative analysis of the existing analogues was executed and the main features of gaze tracking systems were highlighted and prioritized. These features are price, tracking accuracy, angle error, flexibility, and usability. A methodology was developed which allows to calculate a gaze direction vector according to the relative position of eye center and corneal reflection from an infrared diode. The centers of an eye and reflection are estimated using the vector field of image gradients and additional weighting. CUDA technology is used to accelerate the developed algorithms. The main advantage of the developed algorithm is the ability to detect and continuously track pupils’ centers, regardless of the head position, which significantly extends the scope of the gaze tracking system under consideration.

image processing, GPU, AHP, gaze-tracking, human-computer interaction, infrared illumination, CUDA, parallel computing