Head pose tracking for immersive applications

The paper describes a 3D head pose tracking system designed for immersive applications. The proposed system is based on a random forest's head detection and pose regression model. The main novel contributions include the application, in immersive technologies, of a head tracker using a range data sensor, the Kalman filter and the random forest, with automatic detection of outliers and handling of the missing data. The envisaged applications of the proposed system include data normalization and data acquisition from remote locations in real-time. The latter is described in detail by integrating 3 degrees of freedom robotic camera head with the proposed head tracker. In this instance the proposed non-contact head tracking system is used to control the robotic camera by replicating measured operator's head motion to improve his/her spatial awareness and sense of immersion, with the captured video shown on a head mounted display. The experimental section includes accuracy and robustness analysis. The system's robustness is examined with respect to the head pose outliers and the missing head detections. The effects of the presence of the head mounted displays on the performance of the system is also assessed. Additionally a brief discussion of operators' psychophysical perception test is also included.