Abstract:This research has been investigating an automatic and online crowd anomaly detection model by exploring a novel compound image descriptor generated from live video streams. A dual-channel convolutional neural network (DCCNN) has been set up for efficiently processing scene-related and motion-related crowd information inherited from raw frames and the compound descriptor instances. The novelty of the work stemmed from the creation of the spatio-temporal cuboids in online (or near real-time) manner through dynamically extracting local feature tracklets within the temporal space while handling the foreground region-of-interests (i.e., moving targets) through the exploration of Gaussian Mixture Model in the spatial space. Hence, the extracted foreground blocks can effectively eliminate irrelevant backgrounds and noises from the live streams for reducing the computational costs in the subsequent detecting phases. The devised compound feature descriptor, named as spatio-temporal feature descriptor (STFD), is capable of characterizing the crowd attributes through the measures such as collectiveness, stability, conflict and density in each online generated spatio-temporal cuboid. A STFD instance registers not only the dynamic variation of the targeted crowd over time based on local feature tracklets, but also the interaction information of neighborhoods within a crowd, e.g., the interaction force through the K-nearest neighbor (K-NN) analysis. The DCCNN developed in this research enables online identification of suspicious crowd behaviors based on analyzing the live-feed images and their STFD instances. The proposed model has been developed and evaluated against benchmarking techniques and databases. Experimental results have shown substantial improvements in terms of detection accuracy and efficiency for online crowd abnormal behavior identification.

Using Deep Learning Models Combined with Crowd Emotion Models to Identify Abnormal Behaviors in Crowds

Emotion-Based Crowd Representation for Abnormality Detection

Abnormal Behavior Learning Based on Edge Computing toward a Crowd Monitoring System

Emotion Detection in Online Social Networks: A Multilabel Learning Approach

Deep Learning Based Abnormal Behavior Detection for Elderly Healthcare Using Consumer Network Cameras

Dual-channel CNN for efficient abnormal behavior identification through crowd feature engineering

Energy Level-Based Abnormal Crowd Behavior Detection

Detecting and Quantifying Crowd-Level Abnormal Behaviors in Crowd Events

Crowd-level Abnormal Behavior Detection via Multi-scale Motion Consistency Learning

Abnormal crowd behavior detection based on local pressure model

Crowd Counting and Abnormal Behavior Detection via Multiscale GAN Network Combined with Deep Optical Flow

Online anomaly detection in crowd scenes via structure analysis

Extracting macroscopic quantities in crowd behaviour with deep learning

Plug-and-Play CNN for Crowd Motion Analysis: An Application in Abnormal Event Detection

Anomaly Detection and Localisation in the Crowd Scenes Using a Block-Based Social Force Model

Abnormal Behavior Recognition Based on 3D Dense Connections

Neural Networks Assist Crowd Predictions in Discerning the Veracity of Emotional Expressions

Abnormal Behavior Detection Based on Regional Optical Flow

Abnormal behavior recognition using 3D-CNN combined with LSTM

Abnormal Event Detection and Location for Dense Crowds using Repulsive Forces and Sparse Reconstruction

Research on Human Abnormal Behavior Detection Based on Deep Learning