Learning Manifold Patch-Based Representations of Man-Made Shapes

Dmitriy Smirnov, Mikhail Bessmeltsev, Justin Solomon

Keywords: 3D shape representations, CAD modeling, sketch-based modeling, computer graphics, computer vision, deep learning

Abstract: Choosing the right representation for geometry is crucial for making 3D models compatible with existing applications. Focusing on piecewise-smooth man-made shapes, we propose a new representation that is usable in conventional CAD modeling pipelines and can also be learned by deep neural networks. We demonstrate its benefits by applying it to the task of sketch-based modeling. Given a raster image, our system infers a set of parametric surfaces that realize the input in 3D. To capture piecewise smooth geometry, we learn a special shape representation: a deformable parametric template composed of Coons patches. Naively training such a system, however, is hampered by non-manifold artifacts in the parametric shapes and by a lack of data. To address this, we introduce loss functions that bias the network to output non-self-intersecting shapes and implement them as part of a fully self-supervised system, automatically generating both shape templates and synthetic training data. We develop a testbed for sketch-based modeling, demonstrate shape interpolation, and provide comparison to related work.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: We propose a parametrically defined patch-based 3D shape representation that is compatible both with traditional CAD modeling tools and modern deep learning pipelines.

Code: [ dmsm/LearningPatches](https://github.com/dmsm/LearningPatches)

Data: [ShapeNet](https://paperswithcode.com/dataset/shapenet)