PT Unknown
AU Iiris Lusi
   Sergio Escalera
   Gholamreza Anbarjafari
TI Human Head Pose Estimation on SASE database using Random Hough Regression Forests
BT 23rd International Conference on Pattern Recognition Workshops
PY 2016
VL 10165
DI 10.1007/978-3-319-56687-0_12
AB In recent years head pose estimation has become an important task in face analysis scenarios. Given the availability of high resolution 3D sensors, the design of a high resolution head pose database would be beneficial for the community. In this paper, Random Hough Forests are used to estimate 3D head pose and location on a new 3D head database, SASE, which represents the baseline performance on the new data for an upcoming international head pose estimation competition. The data in SASE is acquired with a Microsoft Kinect 2 camera, including the RGB and depth information of 50 subjects with a large sample of head poses, allowing us to test methods for real-life scenarios. We briefly review the database while showing baseline head pose estimation results based on Random Hough Forests.
ER