3D constrained local model-based face recognition using Kinect under variant conditions

This paper presents an algorithm to recognize 3D face under various conditions using 3D constrained local model (CLM-Z). We used a combination of 2D images (RGBs) and depth images (Ds) captured by Kinect which is an inexpensive and affordable sensor. Three-dimensional constrained local model was used for face-modeling and determining the face important points for robust face recognition under challenging conditions. The challenging conditions involved various illuminations, expressions and poses. In addition, we used feature descriptors to obtain feature vectors around each important point (landmark). The proposed method was evaluated with CurtinFaces which is a publicly available dataset. We concluded that the proposed method outperformed the state of the arts methods under various illumination conditions, various expression conditions and pitch pose conditions and comparable results were obtained in other cases.