Generating and understanding the 3D shape of objects is considered as a significant step in many vision applications. This project aims at generating 3D models from 2D images. The architecture of the developed model, employs 3D Vectorization and Generation. It forms 3D objects using the probabilistic space and by taking advantages from novel developments in volumetric convolutional networks and generative adversarial nets. The system uses an adversarial model that is capable of implicitly getting the object description and to produce quality 3D objects and shape descriptors.