Learning a Hierarchical Latent-Variable Model of 3D Shapes
This addresses the challenge of 3D shape understanding and generation for computer vision and graphics applications, though it appears incremental as it builds on existing generative modeling techniques.
The paper tackles the problem of learning hierarchical latent representations of 3D shapes in an unsupervised manner, resulting in a model that improves generalization and performs comparably to state-of-the-art methods on tasks like 3D model retrieval from 2D images.
We propose the Variational Shape Learner (VSL), a generative model that learns the underlying structure of voxelized 3D shapes in an unsupervised fashion. Through the use of skip-connections, our model can successfully learn and infer a latent, hierarchical representation of objects. Furthermore, realistic 3D objects can be easily generated by sampling the VSL's latent probabilistic manifold. We show that our generative model can be trained end-to-end from 2D images to perform single image 3D model retrieval. Experiments show, both quantitatively and qualitatively, the improved generalization of our proposed model over a range of tasks, performing better or comparable to various state-of-the-art alternatives.