A New Volumetric CNN for 3D Object Classification Based on Joint Multiscale Feature and Subvolume Supervised Learning Approaches

Figure 14

Epoch vs loss is computed on 24 views (0.32).