The thesis explores the effectiveness of the Microsoft Kinect sensor for the tracking of human face and estimation of person's gaze point. By using Microsoft Kinect Software Developement Kit we have implemented a face tracking and gaze estimation application, which works as a server providing gaze data for other applications. A web application was developed which visualizes the gaze point and allows measurements of the gaze point accuracy.
The first part of the thesis presents an importance of visual user observations, the types of 2D and 3D sensors, and the Kinect sensor. The second part describes in detail the implementation of the gaze point estimation system using the Kinect SDK. The last part of the thesis presents results of the measurements of gaze accuracy with respect to the distance of the user.
|