摘要：可微渲染推动了新视角合成（NVS）的发展，但由于采集视角稀疏、存在动态物体以及多轨迹数据有限，将其应用于真实世界驾驶场景仍然困难。我们提出多视角多载具（MV2）数据集和基准，用于评估 NVS 模型在动态城市场景中面对大幅视角变化时的表现。MV2 包含来自汽车、踏板车和无人机的同步采集数据，它们分别沿不同但时间同步的轨迹行进。在一辆载具的相机视频流上训练 NVS 方法、并在另一辆载具的数据上进行测试，可实现比现有单轨迹数据集大得多的视角变化评估。所有序列均通过运动恢复结构（Structure-from-Motion）完成配准，相机位姿则借助手工像素级对应标注进行验证，最终形成包含 12,000 张图像的 50 个高质量场景。对近期 NVS 和相机位姿估计方法的基准测试表明，NVS 性能会随视角差异扩大而下降，前馈式位姿估计器也明显落后于基于优化的方法；这凸显了 MV2 作为驾驶场景 NVS 严格测试平台的价值。数据集、基准测试协议和项目资源可通过此 https 链接获取。

Abstract:Differentiable rendering has advanced novel view synthesis (NVS), yet applying it to real-world driving remains difficult due to sparse capture viewpoints, dynamic objects, and limited multi-trajectory data. We introduce the Multi-View Multi-Vehicle (MV2) dataset and benchmark for evaluating NVS models under large viewpoint changes in dynamic urban scenes. MV2 features synchronized captures from a car, scooter, and drone, each following distinct yet synchronized trajectories. Training NVS methods on one vehicle's camera stream and testing on another enables evaluation under substantially larger viewpoint variations than existing single-trajectory datasets. All sequences are registered via Structure-from-Motion and camera poses verified using manual pixel-level correspondence annotations, yielding 50 high-quality scenes with 12000 images. Benchmarking recent NVS and camera pose estimation methods shows that NVS performance degrades with increasing viewpoint disparity, and that feed-forward pose estimators notably lag behind optimization-based approaches, highlighting MV2 as a rigorous testbed for NVS in driving. The dataset, benchmark protocol, and project resources are available at this https URL.