A Novel Hybrid Approach for 3D Face Recognition Based on Higher Order Tensor Mohcene Bessaoudi 1( ✉ ) , Mebarka Belahcene 1 , Abdelmalik Ouamane 1 , Ammar Chouchane 1 , and Salah Bourennane 2 1 LI3C Laboratory, University of Biskra, Biskra, Algeria bessaoudi.mohcene@gmail.com 2 Institut Fresnel, Université de Marseille, Marseille, France Abstract. This paper presents a new hybrid approach for 3D face verification based on tensor representation in the presence of illuminations, expressions and occlusion variations. Depth face images are divided into sub-region and the Multi- Scale Local Binarised Statistical Image Features (MSBSIF) histogram are extracted from each sub-region and arranged as a third order tensor. Furthermore, to reduce the dimensionality of this tensor data, we use a novel hybrid approach based on two steps of dimensionality reduction multilinear and non-linear. Firstly, Multilinear Principal Component Analysis (MPCA) is used. MPCA projects the input tensor in a new lower subspace in which the dimension of each tensor mode is reduced. After that, the non-linear Exponential Discriminant Analysis (EDA) is used to discriminate the faces of different persons. Finally, the matching is performed using distance measurement. The proposed approach (MPCA+EDA) has been tested on the challenging face database Bosporus 3D and the experi‐ mental results show that our method achieves a high verification performance compared with the state of the art. Keywords: 3D face verification · Depth image · Tensor representation · Histograms local features 1 Introduction Face recognition is an important research topic in computer vision area due to their various applications such as, surveillance systems, criminal identification, and human robot-interaction, etc. During the last decades, most research works have been inter‐ esting in two-dimensional images and only a few of them have been utilizing the depth images converted from the 3D scans [1, 2]. 2D face verification and identification are still challenging tasks due to the facial appearance changes caused by many factors in uncontrolled environments such as illuminations, expressions, pose variations, and occlusions. More recently, with the progress of 3D digital sensors and scanners [3], using 3D face information can offer a solution to the previous problems and challenges, and many experiments show that the performance on the 3D shape channel is better than on 2D texture alone [4, 5]. © Springer Nature Switzerland AG 2019 O. Demigha et al. (Eds.): CSA 2018, LNNS 50, pp. 215–224, 2019. https://doi.org/10.1007/978-3-319-98352-3_23