Real-Time Anomaly Detection and Localization in Crowded Scenes
arXiv:1511.06936
Abstract
In this paper, we propose a method for real-time anomaly detection and localization in crowded scenes. Each video is defined as a set of non-overlapping cubic patches, and is described using two local and global descriptors. These descriptors capture the video properties from different aspects. By incorporating simple and cost-effective Gaussian classifiers, we can distinguish normal activities and anomalies in videos. The local and global features are based on structure similarity between adjacent patches and the features learned in an unsupervised way, using a sparse auto- encoder. Experimental results show that our algorithm is comparable to a state-of-the-art procedure on UCSD ped2 and UMN benchmarks, but even more time-efficient. The experiments confirm that our system can reliably detect and localize anomalies as soon as they happen in a video.
arXiv commentCVPRw 2015
Cited by in corpus (6)
- STFCN: Spatio-Temporal FCN for Semantic Video Segmentation
- Learning Memory-guided Normality for Anomaly Detection
- Image/Video Deep Anomaly Detection: A Survey
- Urban Anomaly Analytics: Description, Detection, and Prediction
- Challenges in Time-Stamp Aware Anomaly Detection in Traffic Videos
- Real-Time Anomaly Detection With HMOF Feature