Efficient Part-level 3D Object Generation via Dual Volume Packing

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Tang, Jiaxiang, Lu, Ruijie, Li, Zhaoshuo, Hao, Zekun, Li, Xuan, Wei, Fangyin, Song, Shuran, Zeng, Gang, Liu, Ming-Yu, Lin, Tsung-Yi
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908404643004416
author Tang, Jiaxiang
Lu, Ruijie
Li, Zhaoshuo
Hao, Zekun
Li, Xuan
Wei, Fangyin
Song, Shuran
Zeng, Gang
Liu, Ming-Yu
Lin, Tsung-Yi
author_facet Tang, Jiaxiang
Lu, Ruijie
Li, Zhaoshuo
Hao, Zekun
Li, Xuan
Wei, Fangyin
Song, Shuran
Zeng, Gang
Liu, Ming-Yu
Lin, Tsung-Yi
contents Recent progress in 3D object generation has greatly improved both the quality and efficiency. However, most existing methods generate a single mesh with all parts fused together, which limits the ability to edit or manipulate individual parts. A key challenge is that different objects may have a varying number of parts. To address this, we propose a new end-to-end framework for part-level 3D object generation. Given a single input image, our method generates high-quality 3D objects with an arbitrary number of complete and semantically meaningful parts. We introduce a dual volume packing strategy that organizes all parts into two complementary volumes, allowing for the creation of complete and interleaved parts that assemble into the final object. Experiments show that our model achieves better quality, diversity, and generalization than previous image-based part-level generation methods.
format Preprint
id arxiv_https___arxiv_org_abs_2506_09980
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Efficient Part-level 3D Object Generation via Dual Volume Packing
Tang, Jiaxiang
Lu, Ruijie
Li, Zhaoshuo
Hao, Zekun
Li, Xuan
Wei, Fangyin
Song, Shuran
Zeng, Gang
Liu, Ming-Yu
Lin, Tsung-Yi
Computer Vision and Pattern Recognition
Recent progress in 3D object generation has greatly improved both the quality and efficiency. However, most existing methods generate a single mesh with all parts fused together, which limits the ability to edit or manipulate individual parts. A key challenge is that different objects may have a varying number of parts. To address this, we propose a new end-to-end framework for part-level 3D object generation. Given a single input image, our method generates high-quality 3D objects with an arbitrary number of complete and semantically meaningful parts. We introduce a dual volume packing strategy that organizes all parts into two complementary volumes, allowing for the creation of complete and interleaved parts that assemble into the final object. Experiments show that our model achieves better quality, diversity, and generalization than previous image-based part-level generation methods.
title Efficient Part-level 3D Object Generation via Dual Volume Packing
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.09980