PatchNets: Patch-Based Generalizable Deep Implicit 3D Shape Representations
This work addresses the data efficiency and generalizability challenges in 3D shape modeling for computer vision and graphics applications, offering a more controllable representation.
The paper tackles the problem of requiring large datasets for training deep implicit 3D shape representations by introducing a patch-based method that improves generalizability across object categories and reduces data needs. It shows that training on one category can represent detailed shapes from others and requires fewer shapes compared to existing approaches.
Implicit surface representations, such as signed-distance functions, combined with deep learning have led to impressive models which can represent detailed shapes of objects with arbitrary topology. Since a continuous function is learned, the reconstructions can also be extracted at any arbitrary resolution. However, large datasets such as ShapeNet are required to train such models. In this paper, we present a new mid-level patch-based surface representation. At the level of patches, objects across different categories share similarities, which leads to more generalizable models. We then introduce a novel method to learn this patch-based representation in a canonical space, such that it is as object-agnostic as possible. We show that our representation trained on one category of objects from ShapeNet can also well represent detailed shapes from any other category. In addition, it can be trained using much fewer shapes, compared to existing approaches. We show several applications of our new representation, including shape interpolation and partial point cloud completion. Due to explicit control over positions, orientations and scales of patches, our representation is also more controllable compared to object-level representations, which enables us to deform encoded shapes non-rigidly.