发表机构
Institute of High Performance Computing, Agency for Science, Technology and Research（高性能计算研究所，新加坡科技研究局）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究针对隐私感知课堂事件识别场景，提出轻量鲁棒的运动推理框架，通过知识蒸馏实现高效推理，性能优于大模型且泛化能力强，将公开相关基准与工具。
AI 中文摘要
计算机视觉能否助力提升课堂安全性？在这项试点研究中，我们从类监控摄像头（CCTV-style）观测视角研究隐私感知且计算高效的课堂事件识别任务。该场景探索不足，针对实际部署所需的隐私性、效率和泛化需求的基准及方法均有限。我们引入一种新型混合基准，将生成的类监控摄像头视频与真实课堂姿态数据相结合，并提出一种轻量且鲁棒的运动推理框架，其动机源于观察发现：许多事件的差异更多体现在运动方向、速度、加速度和强度上，而非仅体现在姿态上。为此，我们的方法首先构建人类动作的分层运动学表示，随后将大型教师模型的分层多阶运动学推理知识蒸馏至一个小得多的单阶学生模型，从而实现高效的单人推理，同时保留对运动的表达性理解。实验表明，我们的模型计算成本不到更大基线模型的十分之一，却在性能上大幅优于这些基线，同时展现出更强的域外运动推理能力和零样本合成到真实的泛化能力。我们将公开发布该基准、代码库及配套工具，以推动隐私感知课堂安全领域的进一步研究。
英文摘要
Can computer vision help make classrooms safer? In this pilot study, we investigate privacy-aware and computationally efficient classroom incident recognition from CCTV-style observations. This setting remains underexplored, with limited benchmarks and few methods designed for the privacy, efficiency, and generalization demands of real-world deployment. We introduce a novel hybrid benchmark combining generative CCTV-style videos with real-world classroom pose data, and propose a lightweight, but robust motion-reasoning framework motivated by the observation that many incidents differ more in motion direction, speed, acceleration, and intensity than in pose alone. To that end, our method first constructs hierarchical kinematic representations of human actions. Our method then distills hierarchical, multi-order kinematic reasoning from a large teacher into a much smaller single-order student, enabling efficient per-person inference while preserving expressive motion understanding. Experiments show that our model outperforms substantially larger baselines at less than one-tenth of their computational cost, while also demonstrating stronger out-of-domain motion reasoning and zero-shot synthetic-to-real generalization. We will publicly release the benchmark, codebase, and supporting tools to facilitate further research in privacy-aware classroom safety.
