Enhanced real-time head pose estimation system for mobile device

Article type: Research Article

Authors: Arcoverde Neto, Euclides N.^{a; b; c; *} | Duarte, Rafael M.^b | Barreto, Rafael M.^{a; b} | Magalhães, João Paulo^{a; b} | Bastos, Carlos C.M.^{a; b} | Ren, Tsang Ing^a | Cavalcanti, George D.C.^a

Affiliations: [a] Center for Informatics, Federal University of Pernambuco, Recife, PE, Brazil | [b] C.E.S.A.R – Recife Center for Advanced Studies and Systems, Recife, PE, Brazil | [c] Faculdade Boa Viagem, Recife, PE, Brazil

Correspondence: [*] Corresponding author: Euclides N. Arcoverde Neto, Center for Informatics, Federal University of Pernambuco, Recife, PE, Brazil. E-mail: [email protected]; Site: www.cin.ufpe.br/~viisar.

Abstract: This article proposes a real-time Head Pose Estimation (HPE) technique designed to be used in mobile devices. The method enables the interaction between the user and mobile devices using the device's inbuilt camera. The proposed technique is composed of different computer vision methods, which were optimized to operate in a restricted environment. The method has three Degrees of Freedom (DOF), roll, yaw and pitch. The HPE is obtained using these movements. Experiments were conducted using 363 videos of 27 different people taken in varied scenarios with changes in illumination and background. The results demonstrate the robustness and efficiency of the proposed method.

Keywords: Head pose estimation, adaboost, mobile devices

DOI: 10.3233/ICA-140462

Journal: Integrated Computer-Aided Engineering, vol. 21, no. 3, pp. 281-293, 2014