[论文解读] High-Order Information Matters: Learning Relation and Topology for Occluded Person Re-Identification

Guan’an Wang, Shuo Yang+7 more|arXiv (Cornell University)|Mar 18, 2020

Video Surveillance and Tracking Methods参考文献 42被引用 43

一句话总结

引入一种高阶信息框架用于遮挡重识别，通过自适应定向图卷积建模图像内关系，并通过图匹配实现跨图拓扑，从而产生鲁棒的相似性估计。

ABSTRACT

Occluded person re-identification (ReID) aims to match occluded person images to holistic ones across dis-joint cameras. In this paper, we propose a novel framework by learning high-order relation and topology information for discriminative features and robust alignment. At first, we use a CNN backbone and a key-points estimation model to extract semantic local features. Even so, occluded images still suffer from occlusion and outliers. Then, we view the local features of an image as nodes of a graph and propose an adaptive direction graph convolutional (ADGC)layer to pass relation information between nodes. The proposed ADGC layer can automatically suppress the message-passing of meaningless features by dynamically learning di-rection and degree of linkage. When aligning two groups of local features from two images, we view it as a graph matching problem and propose a cross-graph embedded-alignment (CGEA) layer to jointly learn and embed topology information to local features, and straightly predict similarity score. The proposed CGEA layer not only take full use of alignment learned by graph matching but also re-place sensitive one-to-one matching with a robust soft one. Finally, extensive experiments on occluded, partial, and holistic ReID tasks show the effectiveness of our proposed method. Specifically, our framework significantly outperforms state-of-the-art by6.5%mAP scores on Occluded-Duke dataset.

研究动机与目标

• 将遮挡人群再识别问题定位为存在显著遮挡与错位挑战的问题。
• 提出一种通过图模型学习局部特征之间的高阶关系的框架。
• 提出跨图拓扑模块以在没有硬性一对一匹配的情况下鲁棒对齐两张图像。
• 证明高阶信息在遮挡、部分和整体 ReID 数据集上的显著性能提升。

提出的方法

• 从关键点区域的 CNN 主干与关键点估计模型提取语义局部特征。
• 将局部特征建模为图节点，并应用自适应定向图卷积（ADGC）层来学习并抑制无意义的信息传递。
• 通过跨图嵌入对齐（CGEA）层对两张图像进行对齐，该层通过图匹配嵌入拓扑信息并预测相似性。
• 端到端训练，利用验证损失监督图像对之间的相似性预测。
• 将基于关系的相似性与基于拓扑的相似性融合为最终的标量分数以用于排序。

实验结果

研究问题

• RQ1 高阶关系信息是否可以通过图消息传递修复遮挡区域特征来提高遮挡 ReID 的判别能力？
• RQ2 通过图匹配的跨图拓扑学习是否提供鲁棒的软对齐，在遮挡/部分场景中优于硬性一对一匹配？
• RQ3 提出的 ADGC 与 CGEA 层是否在遮挡、部分和整体 ReID 基准上带来可衡量的收益？
• RQ4 与 state-of-the-art 方法相比，完整的 HOReID 框架在遮挡、部分和整体数据集上的表现如何？

主要发现

• HOReID 在 Occluded-Duke 上超过了当前最先进方法，Rank-1 为 55.1，mAP 为 43.8。
• 在 Occluded-ReID 上，HOReID 实现了 Rank-1 为 80.3，mAP 为 70.2。
• 消融实验表明增加语义特征、高阶关系和拓扑模块的收益（Rank-1：49.9 → 55.1）。
• ADGC 能有效抑制噪声特征，同时提升有意义特征，增强鲁棒性。
• CGEA 实现鲁棒的软跨图对齐，并提升最终相似性预测。