Publication Type : Journal Article
Publisher : Neural Processing Letters
Source : Neural Processing Letters, 53(6), pp.4677-4692
Url : https://link.springer.com/article/10.1007/s11063-021-10618-3
Campus : Chennai
School : School of Computing
Department : Computer Science and Engineering
Year : 2021
Abstract : Real-time video anomaly detection and localization still prevail as a challenging task. Autoencoders are expected to give high reconstruction error for abnormal events than normal events while trained on video segments of normal events. Nevertheless, this assumption is not always true in practice. Sometimes the autoencoder offers better generalization. Therefore, it also reconstructs abnormal events well, leading to slightly degraded performance for anomaly detection. To alleviate this issue, we propose a Skip connected and Memory Guided Network (SMGNet) for video anomaly detection. The memory guided network with skip connection help in avoiding loss of meaningful information such as foreground patterns, in addition to memorizing significant normality patterns. The effect of augmenting memory guided network with skip connection in the residual spatiotemporal autoencoder (R-STAE) architecture is evaluated. The proposed technique achieved improved results over three benchmark datasets.
Cite this Research Publication : Chandrakala, S., Srinivas, V. and Deepak, K., 2021. Residual spatiotemporal autoencoder with skip connected and memory guided network for detecting video anomalies. Neural Processing Letters, 53(6), pp.4677-4692.