MoCA:基于组件的注意力机制用于可扩展的组合3D生成
MoCA: Mixture-of-Components Attention for Scalable Compositional 3D Generation
AI总结:
MoCA通过重要性路由和组件压缩技术,实现高效可扩展的组合性3D生成,优于现有基线方法。
摘要:
组合性对于3D对象和场景生成至关重要,但现有的部分感知3D生成方法在增加组件数量时由于二次全局注意力成本而存在可扩展性差的问题。在本文中,我们提出了MoCA,一种具有两个关键设计的组合性3D生成模型:(1)基于重要性的组件路由,选择最相关的前k个组件进行稀疏全局注意力;(2)不重要组件压缩,保留未选组件的上下文先验,同时减少全局注意力的计算复杂度。通过这些设计,MoCA能够实现高效、精细的组合性3D资产创建,支持可扩展的组件数量。大量实验表明,MoCA在组合性对象和场景生成任务上均优于基线模型。项目页面:https://lizhiqi49.github.io/MoCA
英文摘要:
Compositionality is critical for 3D object and scene generation, but existing part-aware 3D generation methods suffer from poor scalability due to quadratic global attention costs when increasing the number of components. In this work, we present MoCA, a compositional 3D generative model with two key designs: (1) importance-based component routing that selects top-k relevant components for sparse global attention, and (2) unimportant components compression that preserve contextual priors of unselected components while reducing computational complexity of global attention. With these designs, MoCA enables efficient, fine-grained compositional 3D asset creation with scalable number of components. Extensive experiments show MoCA outperforms baselines on both compositional object and scene generation tasks. Project page: https://lizhiqi49.github.io/MoCA
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。