High Order Volumetric Directional Pattern for Video-Based Face Recognition

Figure 5

Honda/UCSD dataset. Each row represents different samples of one subject.