Publication | IEEE International Conference on Computer Vision (ICCV) 2021

LSD-StructureNet

Modeling Levels of Structural Detail in 3D Part Hierarchies

This paper describes a new methodology to generate 3D objects with part hierarchies using neural networks. In particular, it allows conditioning on existing levels, which significantly improves the efficiency on using such approach in interactive scenarios.

Download publication

Abstract

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

Dominic Roberts, Ara Danielyan, Hang Chu, Mani Golparvar-Fard, David Forsyth

IEEE International Conference on Computer Vision (ICCV) 2021

Generative models for 3D shapes represented by hierarchies of parts can generate realistic and diverse sets of out- puts. However, existing models suffer from the key practical limitation of modelling shapes holistically and thus can- not 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 adjust- ing 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.

Related Resources

Publication

2023

Learned Visual Features to Textual Explanations

A novel method that leverages the capabilities of large language…

Publication

2022

MaskTune: Mitigating Spurious Correlations by Forcing to Explore

This work proposes a masking strategy that prevents over-reliance on…

Publication

2021

Cross-Domain Few-Shot Graph Classification

We study the problem of few-shot graph classification across domains…

Publication

2020

Contrastive Multi-View Representation Learning on Graphs

We introduce a self-supervised approach for learning node and graph…

Get in touch

Something pique your interest? Get in touch if you’d like to learn more about Autodesk Research, our projects, people, and potential collaboration opportunities.

Contact us