MER-Tracker：基于多视角事件-RGB 混合相机的高速 3D 点跟踪
MER-Tracker: Towards High-Speed 3D Point Tracking via Multi-View Event-RGB Hybrid Cameras
用事件相机的高时间分辨率弥补 RGB 低帧率，通过线性插值与锚框采样在 3D 提升后融合离散 RGB 特征与连续事件特征，为高速运动下的长时 3D 点轨迹估计提供了可复用的多模态融合范式。
原文依据 apply linear interpolation and anchor sampling to fuse the discrete RGB 3D features and continuous Event 3D features
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论本文提出了首个使用多视角事件-RGB 混合相机进行高速 3D 点跟踪的任务。我们设计了一个由 4 台 RGB 相机（30fps）和 2 台事件相机组成的立方体观测装置，以同步捕捉高速运动，并提出 MER-Tracker，一种融合双模态互补优势的高帧率 3D 点跟踪网络。我们首先分别从 RGB 和事件模态中提取 2D 运动变化特征，随后在 3D 提升之后，采用线性插值与锚框采样来融合离散的 RGB 3D 特征与连续的事件 3D 特征，最后基于时序关联性使用经 LoRA 微调的 Transformer 预测快速运动下的高帧率 3D 点轨迹，从而实现高速 3D 点跟踪。为验证方法的有效性，我们构建了真实世界与仿真的高速运动数据集。在这些数据集上的实验表明，我们的方法能够以高帧率（150fps）实现准确的高速 3D 点跟踪，优于最优方法。
英文原文摘要 1080 字符
This paper proposes the first task for high-speed 3D point tracking using multi-view Event-RGB hybrid cameras. We design a cuboid observation device comprising 4 RGB cameras (30fps) and 2 Event cameras to synchronously capture high-speed motions, and propose MER-Tracker, a high-frame-rate 3D point-tracking network that fuses the complementary strengths of dual modalities. We first respectively extract 2D motion-change features from the RGB and Event modalities, then apply linear interpolation and anchor sampling to fuse the discrete RGB 3D features and continuous Event 3D features after 3D lifting, and finally employ a LoRA-tuned Transformer based on temporal correlationship to predict the high-frame-rate 3D point trajectories over fast motions, accomplishing high-speed 3D point tracking. To verify the effectiveness of our method, we construct both real-world and simulated high-speed motion datasets. Experiments on these datasets show that our method achieves accurate high-speed 3D point tracking at high-frame-rate (150fps), outperforming state-of-the-art methods.