LSD-StructureNet: Modeling Levels of Structural Detail in 3D Part Hierarchies

TL;DR

LSD-StructureNet enhances 3D shape generative models by enabling conditional part sampling within hierarchies, allowing for flexible shape editing at multiple levels without sacrificing realism or speed.

Contribution

It introduces a novel augmentation to StructureNet that allows re-generation of individual parts at arbitrary hierarchy levels through probabilistic conditional decoders.

Findings

Supports conditional sampling without reducing realism or diversity.

Operates efficiently without impacting inference speed.

Validated on the PartNet dataset with positive results.

Abstract

Generative models for 3D shapes represented by hierarchies of parts can generate realistic and diverse sets of outputs. However, existing models suffer from the key practical limitation of modelling shapes holistically and thus cannot perform conditional sampling, i.e. they are not able to generate variants on individual parts of generated shapes without modifying the rest of the shape. This is limiting for applications such as 3D CAD design that involve adjusting created shapes at multiple levels of detail. To address this, we introduce LSD-StructureNet, an augmentation to the StructureNet architecture that enables re-generation of parts situated at arbitrary positions in the hierarchies of its outputs. We achieve this by learning individual, probabilistic conditional decoders for each hierarchy depth. We evaluate LSD-StructureNet on the PartNet dataset, the largest dataset of 3D shapes represented by hierarchies of parts. Our results show that contrarily to existing methods, LSD-StructureNet can perform conditional sampling without impacting inference speed or the realism and diversity of its outputs.