面向位姿图初始化的全局感知边优先排序
Global-Aware Edge Prioritization for Pose Graph Initialization
把位姿图初始化从「逐对检索 top-k 邻居」改为「按边对 SfM 的效用全局排序」，用 GNN 预测全局一致的边可靠性，可直接替换 SfM 前端中基于检索的候选边筛选环节，在稀疏、高速和歧义场景下提升重建鲁棒性。
原文依据 We address this limitation through the concept of edge prioritization, ranking candidate edges by their utility for SfM.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论位姿图是运动恢复结构（SfM）的核心组成部分，其中图像作为节点、边编码相对位姿。由于几何验证代价高昂，SfM 流程将位姿图限制在稀疏的候选边集合上，因此初始化至关重要。现有方法依赖图像检索将每幅图像与其 k 个最近邻相连，独立处理各图像对并忽略全局一致性。我们通过边优先排序的概念解决这一局限，即按候选边对 SfM 的效用对其进行排序。我们的方法包含三个部分：(1) 一个以 SfM 导出的监督训练的 GNN，用于预测全局一致的边可靠性；(2) 由这些排序引导的、基于多最小生成树的位姿图构建；(3) 连通性感知的分数调制，用于强化薄弱区域并减小图直径。这种全局信息驱动的初始化产生更可靠、更紧凑的位姿图，在稀疏和高速场景下提升重建精度，并在歧义场景上优于当前最优的检索方法。代码与训练模型见 https://github.com/weitong8591/global_edge_prior。
英文原文摘要 1174 字符
The pose graph is a core component of Structure-from-Motion (SfM), where images act as nodes and edges encode relative poses. Since geometric verification is expensive, SfM pipelines restrict the pose graph to a sparse set of candidate edges, making initialization critical. Existing methods rely on image retrieval to connect each image to its k nearest neighbors, treating pairs independently and ignoring global consistency. We address this limitation through the concept of edge prioritization, ranking candidate edges by their utility for SfM. Our approach has three components: (1) a GNN trained with SfM-derived supervision to predict globally consistent edge reliability; (2) multi-minimal-spanning-tree-based pose graph construction guided by these ranks; and (3) connectivity-aware score modulation that reinforces weak regions and reduces graph diameter. This globally informed initialization yields more reliable and compact pose graphs, improving reconstruction accuracy in sparse and high-speed settings and outperforming SOTA retrieval methods on ambiguous scenes. The code and trained models are available at https://github.com/weitong8591/global_edge_prior.