用多候选匹配+鲁棒估计提升视觉特征的几何一致性
Preserve, Then Resolve: Many-to-Many Association and Robust Estimation with General-Purpose Visual Features
- 保留多个匹配候选，构建多对多关联图再统一求解
- 在相对位姿估计中，比单候选方法显著提升精度
- 适合依赖视觉特征的3D重建与机器人定位任务
通用视觉特征的语义可迁移性并不保证图像间的几何一致性。我们使用冻结的 DINOv3 特征发现，几何正确的对应点在余弦相似度中常低于排名一，但仍位于一个较小的 top-$K$ 候选集合内。这启发了‘先保留、再求解’的设计：在多对多（m-to-m）关联图中保留多个候选，并将歧义消解推迟到鲁棒估计阶段。我们从概率视角研究 m-to-m 鲁棒估计，将现有匹配基数最大化（MCM）机制解释为似然最大化的主导基数近似，并提出一种更快的实值方法——调和共识最大化（HCM）。两阶段 LO-RANSAC 使用 HCM 进行候选生成，再用 MCM 进行图感知选择。在相对位姿估计任务中，该流程对 DINOv2、DINOv3、V-JEPA 2.1 和 SigLIP 2 均持续优于基线。代码已公开于 https://github.com/LIAS-CUHKSZ/preserve_then_resolve。
原文摘要 · Abstract (English)
The semantic transferability of general-purpose visual features does not guarantee geometric consistency across images. Using frozen DINOv3 features, we show that geometrically correct correspondences often fall below rank one in cosine similarity yet remain within a small top-$K$ candidate set. This motivates a preserve-then-resolve design: we retain multiple candidates in a many-to-many (m-to-m) association graph and defer their disambiguation to robust estimation. We study m-to-m robust estimation from a probabilistic perspective. We interpret the existing Matching Cardinality Maximization (MCM) mechanism as a dominant-cardinality approximation to likelihood maximization and propose a faster, real-valued mechanism called Harmonic Consensus Maximization (HCM). A two-stage LO-RANSAC uses HCM for candidate sourcing and MCM for graph-aware selection. We evaluate end-task gains in relative-pose estimation, where our pipeline consistently improves rank-one baselines for DINOv2, DINOv3, V-JEPA 2.1, and SigLIP 2. Code is available at https://github.com/LIAS-CUHKSZ/preserve_then_resolve.
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。