Anika, AfraKarim, Kazi L.D.Muntaha, RakibunShahrear, FardinAhmed, SupriyoAhmed, Tarem2026-08-302026-08-302017-10-1697815090625532-s2.0-85040000341https://hdl.handle.net/10361/29594This paper focuses on a novel approach for handling radical overhaul of anomalous behavior in a visual surveillance network. The initial objective is online detection of anomalies using a Kernel-based online anomaly detection algorithm. The algorithm will operate onimages collected from a moving camera over a span of space and time. The proposed algorithm established based upon machine learning principles. The eventual target is to connect our proposed system with face detection algorithm. Our proposed system is tested on real data and compared with benchmark existing methods based on the technique of Principal Component Analysis (PCA.) We show that our proposed system achieves high detection accuracy with low computational complexity, while also providing the added benefits of being adaptive, portable, and involving low infrastructural costs.en-USfalseAutomated anomaly detectionFace detectionKernel-based algorithmLearningSensitivityConference Proceeding