Robust Real-Time Unusual Event Detection using Multiple Fixed-Location Monitors

By Amit Adam, Ehud Rivlin, Ilan Shimshoni, David Reinitz. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008

913 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/8a916608e81eea5d7494e577c8563cae44a1b8c6

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Online detection of unusual events in videos via dynamic sparse coding (2011, methodology): “In both videos, there are roughly 10 people walking around in a typical frame, with a frame size of 512 × 384 .T he videos are provided by courtesy of Adam et al. [ 1 ]a nd we comparequantitativelythe detection results of our approach against the method in [10].”
A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction (2021, methodology): “Rather than traditional handcrafted feature based methods [1, 20, 2, 34], a lot of modern deep neural network based methods [11, 45, 50, 31, 25, 30, 8, 36, 38, 47, 43, 24] have been proposed for VAD.”
Deep-anomaly: Fully convolutional neural network for fast anomaly detection in crowded scenes (2016, neutral): “Adam et al. [32] use an exponential distribution for modeling the histograms of optical ﬂow in local regions.”
Anomaly Detection and Localization in Crowded Scenes (2014, methodology): “While simple spatial filtering suffices for some applications [19], crowded scenes require more sophisticated graphical models and inference.”
Observe locally, infer globally: A space-time MRF for detecting abnormal activities with incremental updates (2009, methodology): “Existing approaches vary in the amount of operator supervision entailed, ranging from rule-based approaches [6, 13] to unsupervised methods that directly learn normal activity patterns [19, 18, 2, 1, 16].”
Chaotic invariants of Lagrangian particle trajectories for anomaly detection in crowded scenes (2010, background): “Limiting our review of the literature to work directly related to analysis of crowded scenes, common features used for modeling crowded motions include optical flow [9, 10, 11, 12], gradient [13], spatiotemporal volume [14], etc.”
A Revisit of Sparse Coding Based Anomaly Detection in Stacked RNN Framework (2017, background): “[1] fit the regular histograms of optical flow in local regions with an exponential distribution.”
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].”
Abnormal event detection in crowded scenes using sparse representation (2013, methodology): “Therefore, our MHOF not only describes the motion information as traditional HOF, but also preserves the spatial contextual information .”
Remembering history with convolutional LSTM for anomaly detection (2017, methodology): “In this section, we empirically evaluate our proposed method on a synthesized data (the Moving-MNIST dataset [12]) and some publicly available anomaly detection datasets, including the CUHK Avenue dataset [4] ,the UCSD Pedestrian dataset [20], and the Subway Entrance and Exit dataset [21].”
Self-Supervised Predictive Convolutional Attentive Block for Anomaly Detection (2021, methodology): “Anomaly detection methods can be classified into: dictionary learning methods [7–9, 14, 40, 55], change detection frameworks [11, 26, 38, 48], probability-based methods [1, 2, 16, 23, 29, 44, 45, 56, 61, 74], distance-based models [6, 10, 25, 27, 50, 51,53,57,58,63,65,68,70], and reconstruction-base”
Learning Deep Representations of Appearance and Motion for Anomalous Event Detection (2015, methodology): “Similar to the appearance feature pipeline, the patches are normalized into [0,1] within each channel and 4 encoding layers are used.”
An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos (2018, methodology): “Histogram of oriented optical flow and oriented gradients [30] is a baseline used in anomaly detection and crowd analysis, [31], [32], [33].”