DecompDreamer: A Composition-Aware Curriculum for Structured 3D Asset Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914074029195264 |
|---|---|
| author | Nath, Utkarsh Goel, Rajeev Khurana, Rahul Min, Kyle Ollila, Mark Turaga, Pavan Jampani, Varun Gowda, Tejaswi |
| author_facet | Nath, Utkarsh Goel, Rajeev Khurana, Rahul Min, Kyle Ollila, Mark Turaga, Pavan Jampani, Varun Gowda, Tejaswi |
| contents | Current text-to-3D methods excel at generating single objects but falter on compositional prompts. We argue this failure is fundamental to their optimization schedules, as simultaneous or iterative heuristics predictably collapse under a combinatorial explosion of conflicting gradients, leading to entangled geometry or catastrophic divergence. In this paper, we reframe the core challenge of compositional generation as one of optimization scheduling. We introduce DecompDreamer, a framework built on a novel staged optimization strategy that functions as an implicit curriculum. Our method first establishes a coherent structural scaffold by prioritizing inter-object relationships before shifting to the high-fidelity refinement of individual components. This temporal decoupling of competing objectives provides a robust solution to gradient conflict. Qualitative and quantitative evaluations on diverse compositional prompts demonstrate that DecompDreamer outperforms state-of-the-art methods in fidelity, disentanglement, and spatial coherence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_11981 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DecompDreamer: A Composition-Aware Curriculum for Structured 3D Asset Generation Nath, Utkarsh Goel, Rajeev Khurana, Rahul Min, Kyle Ollila, Mark Turaga, Pavan Jampani, Varun Gowda, Tejaswi Computer Vision and Pattern Recognition Current text-to-3D methods excel at generating single objects but falter on compositional prompts. We argue this failure is fundamental to their optimization schedules, as simultaneous or iterative heuristics predictably collapse under a combinatorial explosion of conflicting gradients, leading to entangled geometry or catastrophic divergence. In this paper, we reframe the core challenge of compositional generation as one of optimization scheduling. We introduce DecompDreamer, a framework built on a novel staged optimization strategy that functions as an implicit curriculum. Our method first establishes a coherent structural scaffold by prioritizing inter-object relationships before shifting to the high-fidelity refinement of individual components. This temporal decoupling of competing objectives provides a robust solution to gradient conflict. Qualitative and quantitative evaluations on diverse compositional prompts demonstrate that DecompDreamer outperforms state-of-the-art methods in fidelity, disentanglement, and spatial coherence. |
| title | DecompDreamer: A Composition-Aware Curriculum for Structured 3D Asset Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.11981 |