Holistic Representation Learning for Multitask Trajectory Anomaly Detection
This work addresses anomaly detection in videos for surveillance or monitoring applications, presenting an incremental improvement over existing methods.
The paper tackled video anomaly detection using skeleton trajectories by proposing a holistic representation learning approach with multitask learning to reconstruct occluded temporal segments, achieving state-of-the-art results on three datasets.
Video anomaly detection deals with the recognition of abnormal events in videos. Apart from the visual signal, video anomaly detection has also been addressed with the use of skeleton sequences. We propose a holistic representation of skeleton trajectories to learn expected motions across segments at different times. Our approach uses multitask learning to reconstruct any continuous unobserved temporal segment of the trajectory allowing the extrapolation of past or future segments and the interpolation of in-between segments. We use an end-to-end attention-based encoder-decoder. We encode temporally occluded trajectories, jointly learn latent representations of the occluded segments, and reconstruct trajectories based on expected motions across different temporal segments. Extensive experiments on three trajectory-based video anomaly detection datasets show the advantages and effectiveness of our approach with state-of-the-art results on anomaly detection in skeleton trajectories.