Privacy-Aware Video Anomaly Detection through Orthogonal Subspace Projection
通过正交子空间投影实现隐私感知的视频异常检测
机构 * School of Engineering and Built Environment, Griffith University(格里菲斯大学工程与环境学院) ; School of Computer Science and Engineering, University of New South Wales(新南威尔士大学计算机科学与工程学院) ; School of Information and Communication Technology, Griffith University(格里菲斯大学信息与通信技术学院)
专题命中 隐私与版权 :alignment(abstract);分类 cs.AI、cs.LG
AI总结 本文提出正交投影层(OPL)和引导OPL(G-OPL),通过去除无关变化提升异常检测相关特征,同时抑制面部属性以保护隐私,实验表明隐私约束能减少敏感信息并保持检测精度。
Comments Accepted as a Spotlight paper at the Forty-Third International Conference on Machine Learning (ICML 2026)