Anomaly Detection in Moving Crowds through Spatiotemporal Autoencoding and Additional Attention

Abstract

We propose an anomaly detection approach by learning a generative model using deep neural network. A weighted convolutional autoencoder- (AE-) long short-term memory (LSTM) network is proposed to reconstruct raw data and perform anomaly detection based on reconstruction errors to resolve the existing challenges of anomaly detection in complicated definitions and background influence. Convolutional AEs and LSTMs are used to encode spatial and temporal variations of input frames, respectively. A weighted Euclidean loss is proposed to enable the network to concentrate on moving foregrounds, thus restraining background influence. Moving foregrounds are segmented from the input frames using robust principal component analysis decomposition. Comparisons with state-of-the-art approaches indicate the superiority of our approach in anomaly detection. Generalization of anomaly detection is improved by enforcing the network to focus on moving foregrounds.

Anomaly Detection in Videos Using Optical Flow and Convolutional Autoencoder

“…The effectiveness of the proposed framework was tested by comparing our model with six different approaches based on autoencoder. These are Con-vAE [12], ST-AE [2], ConvLSTM-AE [40], Two-Stream R-ConvVAE [41] and WCAE-LSTM [25], and STAN [42]. ConvAE [12] benefits from both fully connected autoencoder with trajectory-based handcrafted spatio-temporal features and convolutional autoencoder.…”

“…Two-Stream R-ConvVAE [41] aimed to model regular scenes by using two-stream recurrent variational autoencoder in a semisupervised learning manner. WCAE-LSTM [25] proposed a weighted ConvAE and LSTM in order to encode spatial and temporal information of video frames. WCAE-LSTM [25] proposed a weighted Euclidean loss in order to extract foreground objects.…”

“…Thus, the framework allows the extraction of spatial features and their temporal evolutions. In a similar study [25], The ConvAE + ConvLSTM model was trained by minimizing a Weighted Euclidean Loss. Principal Component Analysis (PCA) was used for segmentation of moving foregrounds from the video volume, and then the loss function was calculated with only extracted foregrounds.…”

“…ConvLSTM [34] allows LSTM to work on video data by substituting the matrix multiplication of LSTM with the convolutional operation. In this way, ConvLSTM captures both temporal and spatial information and provides superior performance with video frames comparing to fully connected LSTM [25]. As shown in Fig.…”

Violence Detection In Surveillance Videos Using Deep Learning

“…Yang et al [7] propose an anomaly discovery approach by learning a generative model utilizing deep neural network system a weighted convolutional autoencoder-(AE-) long short term memory (LSTM) network is proposed to reconstruct data and perform anomaly location dependent on reconstruction errors to determine the current difficulties of irregularity recognition in complicated definitions and background impact. this methodology accomplished an accuracy of 85.7% on CUHK Avenue Dataset ,85.1% on UCSD Ped1 dataset and 92.6% on Ped2 Dataset.…”