RaGS：释放 3D 高斯泼溅，融合 4D 雷达与单目线索进行 3D 目标检测
RaGS: Unleashing 3D Gaussian Splatting from 4D Radar and Monocular Cue for 3D Object Detection
将场景建模为连续高斯场，用视锥初始化与多模态迭代聚合动态分配资源到前景区域，再渲染为层级 BEV 特征；这种以高斯为中间表示、按需聚焦稀疏信息区域的做法，可迁移到其他多传感器 3D 检测与稀疏场景感知任务。
原文依据 models the scene as a continuous field of Gaussians, enabling dynamic resource allocation to foreground objects detection
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论4D 毫米波雷达是自动驾驶中一种有前景的传感模态，然而基于 4D 雷达与单目图像的有效 3D 目标检测仍具挑战性。现有融合方法要么依赖缺乏全局上下文的实例提案，要么依赖受刚性结构约束的稠密鸟瞰图网格，缺乏对多样场景灵活自适应的表示。为此，我们提出 RaGS，这是首个将场景建模为连续高斯场的框架，能够动态分配资源以检测前景目标，同时保持灵活性与效率。具体而言，RaGS 采用级联流水线来构建并逐步精化高斯场。它首先通过基于视锥的定位初始化（FLI）将前景像素反投影以初始化粗略的高斯中心。然后，迭代多模态聚合（IMA）显式利用图像语义并隐式整合 4D 雷达速度几何，以在感兴趣区域内精化高斯。最后，多级高斯融合（MGF）将高斯场渲染为层级鸟瞰图特征，用于 3D 目标检测。通过动态聚焦于稀疏且信息丰富的区域，RaGS 实现了以目标为中心的精度与全面的场景感知。在 View-of-Delft、TJ4DRadSet 和 OmniHD-Scenes 上的大量实验证明了其鲁棒性与最先进性能。源代码见 https://github.com/shawnnnkb/RaGS。
英文原文摘要 1466 字符
4D millimeter-wave radar is a promising sensing modality for autonomous driving, yet effective 3D object detection from 4D radar and monocular images remains challenging. Existing fusion approaches either rely on instance proposals lacking global context or dense BEV grids constrained by rigid structures, lacking a flexible and adaptive representation for diverse scenes. To address this, we propose RaGS, the first framework that models the scene as a continuous field of Gaussians, enabling dynamic resource allocation to foreground objects detection while maintaining flexibility and efficiency. Specifically, RaGS adopts a cascaded pipeline to construct and progressively refine the Gaussian field. It begins with Frustum-based Localization Initiation (FLI), which unprojects foreground pixels to initialize coarse Gaussian centers. Then, Iterative Multimodal Aggregation (IMA) explicitly exploits image semantics and implicitly integrates 4D radar velocity geometry to refine the Gaussians within regions of interest. Finally, Multi-level Gaussian Fusion (MGF) renders the Gaussian field into hierarchical BEV features for 3D object detection. By dynamically focusing on sparse and informative regions, RaGS achieves object-centric precision and comprehensive scene perception. Extensive experiments on View-of-Delft, TJ4DRadSet, and OmniHD-Scenes demonstrate its robustness and SOTA performance. Source code is available at https://github.com/shawnnnkb/RaGS.