RoMa v2提升三维图像密集匹配精度与速度，更适应复杂现实场景。
RoMa v2: Harder Better Faster Denser Feature Matching
- 设计新架构与损失函数，结合多样化训练数据提升匹配能力。
- 在MegaDepth和ScanNet上准确率显著超越前代模型，达90.3%与88.7%。
- 适合需要高精度、低延迟的三维视觉任务，如机器人导航与重建。
密集特征匹配旨在估计三维场景中两幅图像之间的所有对应关系，因其高精度与鲁棒性已成为行业标准。然而，现有密集匹配器在许多真实复杂场景下仍表现不佳，且高精度模型通常速度缓慢，限制了实际应用。本文通过一系列系统性改进，全面攻克上述问题。我们提出新型匹配架构与损失函数，结合精心筛选的多样化训练分布，使模型能解决诸多复杂匹配任务。进一步采用解耦的两阶段匹配-精修流水线加速训练，并通过自定义CUDA内核显著降低精修阶段内存消耗。最后，融合DINOv3基础模型及其他优化策略，增强模型鲁棒性与公平性。在大量实验中，新匹配器达到新的状态最优水平，显著优于以往方法。代码已开源：https://github.com/Parskatt/romav2。
原文摘要 · Abstract (English)
Dense feature matching aims to estimate all correspondences between two images of a 3D scene and has recently been established as the gold standard due to its high accuracy and robustness. However, existing dense matchers still fail or perform poorly for many hard real-world scenarios, and high-precision models are often slow, limiting their applicability. In this paper, we attack these weaknesses on a wide front through a series of systematic improvements that together yield a significantly better model. In particular, we construct a novel matching architecture and loss, which, combined with a curated diverse training distribution, enables our model to solve many complex matching tasks. We further make training faster through a decoupled two-stage matching-then-refinement pipeline, and at the same time, significantly reduce refinement memory usage through a custom CUDA kernel. Finally, we leverage the recent DINOv3 foundation model along with multiple other insights to make the model more robust and unbiased. In our extensive set of experiments, we show that the resulting novel matcher sets a new state-of-the-art, being significantly more accurate than its predecessors. Code is available at https://github.com/Parskatt/romav2
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。