Vision-based head tracking method for eye-controlled human/computer interface

Abstract

As computers become more powerful, the critical bottleneck in their use is often in the interface to the user rather than the computing power. One of the goals in human computer interaction (HCI) research is to increase the communication bandwidth between the user and the computer. Due to recent technological advances, a number of novel HCI modalities have emerged?hand/head gestures, speech, eye tracking, etc. The aim of this study is to develop and evaluate a simple vision-based head tracking for eye-controlled human/computer interface. Tracking head movements is important in the design of an eye-controlled human/computer interface. Two vision-based head tracking systems are proposed to allow the user’s head movements in the design of the eye-controlled human/computer interface. In the first part, we present a camera-mounted head tracking method. A vision-based head tracking method which is robust to electromagnetic interference and ferrometallic objects is designed. We propose an efficient camera calibration method for providing the accurate head pose. We also address the relation between the accuracy of camera parameters and the configuration of calibration points. In the second part, we present a marker-attached head tracking method. A vision-based head tracking method which is simple and non-intrusive is designed. We calculate the position and orientation of the user’s head using the perspective- 3-points problem, and select one reasonable solution among the solution candidates using the property of head movements. In the third part, we describe two experiments which are performed to assess the performance of the two vision-based head tracking methods. A few key measures of performance such as accuracy, stability, and data rate are used. To assess the influence of the configuration of calibration points, four different configurations are investigated in Experiment 1. We compare the performance of the camera-mounted head tracking method (CMM) with that of ...