Abstract:The recognition of abnormal behavior in surveillance video is the focus of current research, which has high research value and broad application possibilities. Its main applications are in the fields of intelligent surveillance, intelligent security, and smart cities, and it is of great significance to study the recognition of abnormal behaviors. Because of the complexity of human movement and the variability of the external environment, the recognition and detection of abnormal behaviors have some challenges. The recognition and detection of abnormal human behaviors in surveillance video still needs further research and development. This paper uses the multi-branch convolutional neural network to extract the spatial features of video frames for the first time, and as an encoder to pass the condensed features to the Gated Recurrent Unit (GRU), which extracts Temporal features from multiple video frames. And then the Gated Recurrent Unit output the result as the decoder. We did a series of comparative experiments on UCF-Crime dataset. And finally, we achieved an accuracy of 86.78% in the test set. The experimental results show that our multi-branch convolutional fusion neural network is better than previous surveillance video abnormal behavior recognition algorithms. At the same time, in order to verify the generalization performance and efficiency of the algorithm, we also conducted an experimental validation on the UCF-101 dataset in this paper, and the results show that the algorithm in this paper can also show a high accuracy rate on the UCF-101 dataset, and the speed of the algorithm is almost close to that of the C3D method with improved accuracy rate, making it possible to develop simple recognition applications based on the algorithm studied in this paper subsequently.

An Abnormal Behavior Recognition Method Based on Fusion Features

Abnormal behavior recognition using 3D-CNN combined with LSTM

A study on attention-based LSTM for abnormal behavior recognition with variable pooling

Abnormal Behavior Recognition Based on 3D Dense Connections

Abnormal Activity Recognition Based on Deep Learning in Crowd

Research on Human Abnormal Behavior Detection Based on Deep Learning

A New Method of Abnormal Behavior Detection Using LSTM Network with Temporal Attention Mechanism

Skeleton Based Abnormal Behavior Recognition Using Spatio-Temporal Convolution and Attention-Based LSTM

Improved human action recognition approach based on two-stream convolutional neural network model

Abnormal Behavior Recognition Based on Spatio-temporal Context

Human abnormal behavior recognition based on two-stream convolutional neural network

Abnormal behavior detection algorithm based on multi-branch convolutional fusion neural network

Unusual Human Behavior Detection System in Real-Time Video Systems

Human Action Recognition Based on Three-Stream Network with Frame Sequence Features

Long Short‐term Memory and Convolutional Neural Network for Abnormal Driving Behaviour Recognition

Behavior Recognition Algorithm Based on the Fusion of SE-R3D and LSTM Network

Research into the Applications of a Multi-Scale Feature Fusion Model in the Recognition of Abnormal Human Behavior

Abnormal Human Activity Recognition using Bayes Classifier and Convolutional Neural Network

Abnormal Behavior Detection of ATM Surveillance Videos Based on Pseudo-3D Residual Network

3D-Cnn-based Fused Feature Maps with LSTM Applied to Action Recognition.

Find the Unseen Actions: Abnormal Action Recognition