# MVCNN - Multi view CNN for 3D object recognition - Limitations - Generating multi-views is time consuming process - Objects are partially visible due to (self) occlusion - Number of categories should be defined in advance. -