A Real-Time Deep Learning Approach for Real-World Video Anomaly Detection

Anomaly detection in video streams with imbalanced data and real-time constraints is a challenging task of computer vision. This paper proposes a novel real-time approach for real-world video anomaly detection exploiting a supervised learning methodology. In particular, we present a deep learning architecture based on the analysis of contextual, spatial, and motion information extracted from the video. A data balancing strategy based on hard-mining and adaptive framerate is used to avoid overfitting and increase detection accuracy. The approach defines an extended taxonomy by differentiating anomalies in "soft" and "hard". A novel anomaly detection score based on a sigmoidal function has been introduced to reduce false positive rate while maintaining a high level of true positive rate. The proposed methodology has been validated with a set of experiments on a well-known video anomaly dataset: UCF-CRIME. The experiments on the testbed demonstrate the impact of the contextual information and data balancing on the classification performances, considering only "hard" anomalies during training and that the proposed model can achieve state-of-the-art performances while minimizing resource consumption.