Viet-Tuan Le (Felix)

Video anomaly detection aims to identify sporadic abnormal events among abundant normal ones in surveillance videos. For this purpose, many existing methods leverage U-Net architecture to predict or reconstruct frames, typically extracting spatial features within a sequence of previous consecutive frames. Yet, several approaches utilize motion information by introducing a two-stream network where the spatial and temporal streams flow independently, fusing the final outputs from these streams to obtain the anomaly score. Given a target optical flow and a sequence of previous frames, it would be desirable to design a framework that predicts the target frame with the intended motion corresponding to the provided optical flow. We present a novel motion-guided prediction framework that uses a proposed motion module to guide the prediction process for generating the output, that is close to its ground truth. Extensive evaluation, including ablation studies, suggests that the proposed approach outperforms the state-of-the-art on three standard benchmark datasets, i.e. UCSD Ped2, CUHK Avenue, and ShanghaiTech. This results suggest that it has the potential for real applications. The code and model are available at https://moguprediction.github.io.

Anomaly detection is to identify abnormal events against normal ones within surveillance videos mainly collected in ground-based settings. Recently, the demand for processing drone-collected data is rapidly growing with the expanding range of drone applications. However, as most aerial videos collected by flying drones contain dynamic backgrounds and others, it is necessary to deal with their spatio-temporal features in detecting anomalies. This study presents a transformer-based video anomaly detection method whereby we investigate a challenging aerial dataset and three benchmark ground-based datasets. A multi-stage transformer is leveraged as an encoder to generate multi-scale feature maps, which are then conveyed to a hierarchical spatio-temporal transformer, that is linked to a decoder and used to capture spatial and temporal information by utilizing a joint attention mechanism. Extensive evaluations including several ablation studies suggest that this network outperforms the state-of-the-art methods. We expect the proposed transformer for U-net can find diverse applications in the video processing area. Code and pre-trained models are publicly available at https://github.com/vt-le/HSTforU.

Automatic anomaly detection is a crucial task in video surveillance system intensively used for public safety and others. The present system adopts a spatial branch and a temporal branch in a unified network that exploits both spatial and temporal information effectively. The network has a residual autoencoder architecture, consisting of a deep convolutional neural network-based encoder and a multi-stage channel attention-based decoder, trained in an unsupervised manner. The temporal shift method is used for exploiting the temporal feature, whereas the contextual dependency is extracted by channel attention modules. System performance is evaluated using three standard benchmark datasets. Result suggests that our network outperforms the state-of-the-art methods, achieving 97.4% for UCSD Ped2, 86.7% for CUHK Avenue, and 73.6% for ShanghaiTech dataset in term of Area Under Curve, respectively.