3DZip通过空间感知的特征多样性压缩，让3D问答模型更快更省资源。
3DZip: Spatial-Aware Feature Diversity-Guided Token Compression for 3D Question Answering
- 先用体素化去冗余，再选特征多样性的关键点，最后按空间约束合并。
- 仅用128个令牌就保留94.7%性能，推理速度提升1.92倍。
- 适合需要高效3D视觉语言理解的场景，如机器人导航与智能交互。
近期3D视觉语言模型通过将2D视觉特征投影到世界坐标系生成几何感知的令牌，支持3D问答等任务中的空间推理。然而，该方法每场景产生数千个令牌，带来巨大计算与内存开销。尽管2D VLMs中已有大量令牌压缩研究，现有方法依赖语义相关性或注意力选择，忽视3D令牌的结构化空间特性。此外，仅靠空间邻近无法解决3D表示中的冗余问题，因物体级令牌分布不均现象在空间聚合后依然存在。为此，我们提出3DZip，一个三阶段令牌压缩框架：首先粗粒度体素化消除点级冗余；其次基于特征空间多样性，利用行列式点过程选取锚点令牌；最后在空间约束下合并剩余令牌，保持几何一致性。在三个3D问答基准上的实验表明，3DZip持续优于现有压缩方法，在仅使用128个令牌的情况下保留94.7%原始性能，实现1.92倍的推理加速。
原文摘要 · Abstract (English)
Recent 3D vision-language models (3D VLMs) construct geometry aware tokens by projecting 2D visual features into world coordinates, enabling spatial reasoning for tasks such as 3D question answering. However, this design generates thousands of tokens per scene, resulting in substantial computational and memory overhead. While token compression has been extensively studied in 2D VLMs, existing approaches rely on semantic relevance or attention-based selection that overlook the structured spatial nature of 3D tokens. Moreover, redundancy in 3D representations cannot be resolved by spatial proximity alone, as object-level token imbalance persists even after spatial aggregation. To address this, we propose 3DZip, a three-stage token compression framework that first applies coarse voxelization to remove point-level redundancy, then selects anchor tokens based on feature-space diversity via a Determinantal Point Process, and finally merges remaining tokens under spatial constraints to preserve geometric coherence. Experiments on three 3D question answering benchmarks demonstrate that 3DZip consistently outperforms existing compression methods, retaining 94.7% of the original performance with only 128 tokens, achieving a $1.92\times$ faster inference speed.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。