Detecting unusual activity in video

By Hua Zhong, Jianbo Shi, Mirkó Visontai. Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004., 2004

563 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/1037b49ca82ad5e4398f3bdf970e93cdf8c7ee81

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A system for learning statistical motion patterns (2006, background): “Zhong et al. [52] detect anomalies with comparison between behaviors but without learned motion patterns.”
Chaotic invariants of Lagrangian particle trajectories for anomaly detection in crowded scenes (2010, background): “Throughout the literature, an anomaly is treated as a context-sensitive term, which is usually studied according to the scenarios of interest [4, 5].”
Machine Recognition of Human Activities: A Survey (2008, methodology): “Filtering with the Gaussian kernel in space and the derivative of the Gaussian on the temporal axis followed by thresholding of the responses and accumulation into spatial histograms was found to be a simple yet effective feature for actions in far-field settings [13].”
Beyond Actions: Discriminative Models for Contextual Group Activities (2010, unknown): “[52] examine motion and shape features of entire video frames to detect unusual activities.”
Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words (2008, background): “Zhong et al. (2004) have proposed an unsupervised approach to detect unusual activity in video sequences.”
Video-Based Abnormal Human Behavior Recognition—A Review (2012, background): “” Such unusual activities are by nature rare, unexpected, atypical [5], and out-of-the-ordinary [4], [70], [71].”
Modelling Crowd Scenes for Event Detection (2006, unknown): “However, unsupervised learning techniques provide the means to learn the typical labels (space-time behavioural patterns) and have been applied for similar problems in video analysis [15] [12].”
Observe locally, infer globally: A space-time MRF for detecting abnormal activities with incremental updates (2009, background): “The above methods [19, 18, 16, 11, 8] can be considered “global”, in the sense that they typically attempt to find the abnormal global activity patterns in a video clip in which several local activities can co-occur.”
Topic models for scene analysis and abnormality detection (2009, background): “in [13] build a database of spatio-temporal patches using normal behavior and detect those patterns that cannot be composed from the database as being abnormal.”
Hierarchical Aligned Cluster Analysis for Temporal Clustering of Human Motion (2013, background): “In the computer vision literature, Zhong et al. [23] used a bipartite graph co-clustering algorithm to segment and detect unusual activities in video.”
Discriminative Latent Models for Recognizing Contextual Group Activities (2012, background): “Most of the work on context is in scene and object recognition.”
A survey on activity recognition and behavior understanding in video surveillance (2013, background): “Automatic recognition of anomalies in a camera’s field of view is one such problem that has attracted attention from vision researchers [190, 229].”
Sparse reconstruction cost for abnormal event detection (2011, background): “[29] and 3D spatio-temporal foreground mask feature fusing Markov Random Field by Benezeth et al.”
Robust Real-Time Unusual Event Detection using Multiple Fixed-Location Monitors (2008, background): “In contrast with most previous works reported, our algorithm satisfies all of these requirements (having of course its own limitations which will be discussed).”
Unsupervised Activity Perception by Hierarchical Bayesian Models (2007, methodology): “The second approach [15] [10] directly uses motion feature vectors instead of tracks to describe video clips.”
Detecting Irregularities in Images and in Video (2005, methodology): “Nevertheless, the representations employed in previous methods have * Patent Pending been either very restrictive (e.g., trajectories of moving objects (Stauffer and Grimson, 2000)), or else too global (e.g., a single small descriptor vector for an entire frame (Zhong et al., 2004)).”
Video Behavior Profiling for Anomaly Detection (2008, background): “Alternatively, the video can be simply sliced into overlapping segments with a fixed time duration [34].”
Anomaly detection in extremely crowded scenes using spatio-temporal motion pattern models (2009, background): “Deviation approaches model usual activity, detecting unusual events as those that differ from the pre-trained model [1, 3, 5, 8, 9, 13, 21, 23, 24, 25].”
A Data-Driven Approach for Event Prediction (2010, methodology): “In this case, the algorithm is given a large corpus of training video captured at a particular location as input, and the objective is to identify abnormal events taking place in the future in that same scene [7–10].”