Performance Evaluation of Head Pose Estimation Methods for Digital Signage Systems
Sunghyun Nam, Gunhee Oh, Myoungho Jang, Heoncheol Lee · 2021 21st International Conference on Control, Automation and Systems (ICCAS) · 2021
This paper addresses the problem of head pose estimation (HPE) for digital signage systems. Because the HPE which estimates the roll, pitch and yaw angles of human heads can extract not only the existence of human faces but also additional information, it is expected that the HPE can play an important role in digital signage systems. However, for the quality of the additional information, the HPE needs to be accurately conducted. This paper presents various performance evaluation results of the HPE methods which are non-deep learning-based and deep learning-based methods. The HPE methods in this work are implemented based on the existing open-sources. Their accuracy was evaluated with the public datasets by comparing the estimated angles and the true angles in the public datasets.