发现并标注了580个长期人类活动周期流程，填补了该领域空白。
Unsupervised Discovery of Long-Term Spatiotemporal Periodic Workflows in Human Activities
- 构建首个包含580个多模态序列的长期周期流程基准数据集
- 提出无需训练的轻量级基线，在三项任务上均显著优于现有方法
- 适用于制造业、体育等场景，免标注可部署，适合实际应用
具有隐式工作流的周期性人类活动在制造、体育和日常生活中普遍存在。尽管短期周期活动（结构简单、模式对比度高）已被广泛研究，但长期周期工作流（模式对比度低）仍基本未被探索。为此，我们提出了首个包含580个多模态人类活动序列的基准数据集，支持三项与实际应用对齐的评估任务：无监督周期工作流检测、任务完成追踪和流程异常检测。我们还提出一种轻量级、无需训练的基线模型，用于建模多样化的周期工作流模式。实验表明：(i) 该基准对现有的无监督周期检测方法和基于大语言模型的零样本方法构成显著挑战；(ii) 我们的基线在所有评估任务中均显著优于对比方法；(iii) 在真实应用中，该基线在部署性能上可媲美传统有监督工作流检测方法，且无需标注与重训练。项目主页：https://sites.google.com/view/periodicworkflow。
原文摘要 · Abstract (English)
Periodic human activities with implicit workflows are common in manufacturing, sports, and daily life. While short-term periodic activities -- characterized by simple structures and high-contrast patterns -- have been widely studied, long-term periodic workflows with low-contrast patterns remain largely underexplored. To bridge this gap, we introduce the first benchmark comprising 580 multimodal human activity sequences featuring long-term periodic workflows. The benchmark supports three evaluation tasks aligned with real-world applications: unsupervised periodic workflow detection, task completion tracking, and procedural anomaly detection. We also propose a lightweight, training-free baseline for modeling diverse periodic workflow patterns. Experiments show that: (i) our benchmark presents significant challenges to both unsupervised periodic detection methods and zero-shot approaches based on powerful large language models (LLMs); (ii) our baseline outperforms competing methods by a substantial margin in all evaluation tasks; and (iii) in real-world applications, our baseline demonstrates deployment advantages on par with traditional supervised workflow detection approaches, eliminating the need for annotation and retraining. Our project page is https://sites.google.com/view/periodicworkflow.
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。