Cross-Modality 2D-3D Face Recognition via Multiview Smooth Discriminant Analysis Based on ELM

Figure 2

The principle of Multiview Smooth Discriminant Analysis (MSDA). The goal of MSDA is to build a smoothing common discriminative space into which the face samples from different views are projected. The different classes of samples are distributed in different views and the MSDA projects the samples into a common subspace to make the samples closer. Here, images with different labels are denoted in different shapes, such as triangle, circle, and square.