通过亲和度对比学习提升骨骼动作识别的区分能力
Affinity Contrastive Learning for Skeleton-based Human Activity Understanding
- 引入亲和度度量构建活动超类，增强对比信号
- 动态温度调节使不同超类惩罚更合理，提升分类性能
- 适合骨骼动作识别、步态分析与行人重识别任务
基于骨骼的人体动作理解中，现有方法多采用对比学习构建判别性特征空间，但常忽略类别间的结构相似性，并受异常正样本干扰。本文提出ACLNet，通过探索动作类别间的复杂聚类关系，提升特征区分度。具体地，设计亲和度度量以优化相似性计算，形成更具信息量的活动超类；引入动态温度调度机制，自适应调整各超类的惩罚强度；并采用基于边距的对比策略，强化类内难分正负样本的分离。在NTU RGB+D 60、NTU RGB+D 120、Kinetics-Skeleton、PKU-MMD、FineGYM和CASIA-B六个数据集上的大量实验表明，该方法在动作识别、步态识别与行人重识别任务上均表现优越。代码已开源。
原文摘要 · Abstract (English)
In skeleton-based human activity understanding, existing methods often adopt the contrastive learning paradigm to construct a discriminative feature space. However, many of these approaches fail to exploit the structural inter-class similarities and overlook the impact of anomalous positive samples. In this study, we introduce ACLNet, an Affinity Contrastive Learning Network that explores the intricate clustering relationships among human activity classes to improve feature discrimination. Specifically, we propose an affinity metric to refine similarity measurements, thereby forming activity superclasses that provide more informative contrastive signals. A dynamic temperature schedule is also introduced to adaptively adjust the penalty strength for various superclasses. In addition, we employ a margin-based contrastive strategy to improve the separation of hard positive and negative samples within classes. Extensive experiments on NTU RGB+D 60, NTU RGB+D 120, Kinetics-Skeleton, PKU-MMD, FineGYM, and CASIA-B demonstrate the superiority of our method in skeleton-based action recognition, gait recognition, and person re-identification. The source code is available at https://github.com/firework8/ACLNet.
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。