Hierarchical Transformers for Unsupervised 3D Shape Abstraction
This addresses the problem of flexible and unsupervised hierarchical representation for 3D shapes, which is incremental as it builds on prior hierarchical methods but removes fixed structural constraints.
The paper tackles unsupervised 3D shape abstraction by introducing HiT, a hierarchical neural field representation that learns general hierarchies across shape categories without supervision, achieving successful multi-level segmentation on all 55 ShapeNet categories.
We introduce HiT, a novel hierarchical neural field representation for 3D shapes that learns general hierarchies in a coarse-to-fine manner across different shape categories in an unsupervised setting. Our key contribution is a hierarchical transformer (HiT), where each level learns parent-child relationships of the tree hierarchy using a compressed codebook. This codebook enables the network to automatically identify common substructures across potentially diverse shape categories. Unlike previous works that constrain the task to a fixed hierarchical structure (e.g., binary), we impose no such restriction, except for limiting the total number of nodes at each tree level. This flexibility allows our method to infer the hierarchical structure directly from data, over multiple shape categories, and representing more general and complex hierarchies than prior approaches. When trained at scale with a reconstruction loss, our model captures meaningful containment relationships between parent and child nodes. We demonstrate its effectiveness through an unsupervised shape segmentation task over all 55 ShapeNet categories, where our method successfully segments shapes into multiple levels of granularity.