Manifold learning techniques for unsupervised anomaly detection
作者：
Highlights：
• Manifold learning techniques are used to construct improved image background models.
• Random uniform sampling of disjoint image neighborhoods yields background sample.
• Detection statistic is distance of remaining neighborhoods from background manifold.
• Performance versus parameters like kernel bandwidth and sampling density is tested.
• Kernel PCA beats diffusion map and benchmark RX on maritime anomaly detection task.
摘要
•Manifold learning techniques are used to construct improved image background models.•Random uniform sampling of disjoint image neighborhoods yields background sample.•Detection statistic is distance of remaining neighborhoods from background manifold.•Performance versus parameters like kernel bandwidth and sampling density is tested.•Kernel PCA beats diffusion map and benchmark RX on maritime anomaly detection task.
论文关键词：Manifolds,Manifold learning,Image processing,Anomaly detection,Target detection
论文评审过程：Received 25 July 2016, Revised 31 July 2017, Accepted 1 August 2017, Available online 25 August 2017, Version of Record 5 October 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.08.005
