Zero-Shot Multi-Object Scene Completion

We present a 3D scene completion method that recovers the complete geometry of multiple unseen objects in complex scenes from a single RGB-D image. Despite notable advancements in single-object 3D shape completion, high-quality reconstructions in highly cluttered real-world multi-object scenes remai...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-08
Main Authors: Iwase, Shun, Liu, Katherine, Guizilini, Vitor, Gaidon, Adrien, Kitani, Kris, Ambrus, Rares, Zakharov, Sergey
Format: Article
Language:English
Subjects:
Datasets
Geometric reasoning
Occlusion
Octrees
Three dimensional models
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a 3D scene completion method that recovers the complete geometry of multiple unseen objects in complex scenes from a single RGB-D image. Despite notable advancements in single-object 3D shape completion, high-quality reconstructions in highly cluttered real-world multi-object scenes remains a challenge. To address this issue, we propose OctMAE, an architecture that leverages an Octree U-Net and a latent 3D MAE to achieve high-quality and near real-time multi-object scene completion through both local and global geometric reasoning. Because a naive 3D MAE can be computationally intractable and memory intensive even in the latent space, we introduce a novel occlusion masking strategy and adopt 3D rotary embeddings, which significantly improves the runtime and scene completion quality. To generalize to a wide range of objects in diverse scenes, we create a large-scale photorealistic dataset, featuring a diverse set of 12K 3D object models from the Objaverse dataset which are rendered in multi-object scenes with physics-based positioning. Our method outperforms the current state-of-the-art on both synthetic and real-world datasets and demonstrates a strong zero-shot capability.
ISSN:2331-8422