A Self-Reasoning Framework for Anomaly Detection Using Video-Level Labels

Zaheer, Muhammad Zaigham; Mahmood, Arif; Shin, Hyoseop; Lee, Seung-Ik

doi:10.1109/lsp.2020.3025688

Cited by 72 publications

(17 citation statements)

References 27 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Weakly-supervised anomaly detection has been studied extensively in the past few years [11,16,13,15,12,17,7,14,4]. Major previous work can be divided into two categories based on the learning paradigm: (1) Multiple Instance Learning (MIL) based approach, (2) Cleaning Noisy Labels based approach.…”

Section: Related Workmentioning

confidence: 99%

“…However, the generation of pseudo temporal annotations is done by training a Graph Convolution Network (GCN). Since training of GCN is computationally complex and can lead to unconstrained latent space, authors in [14] use clustering algorithms for cleaning the noisy labels of untrimmed anomaly videos. Different from [16], Zaheer et al [14] uses the k-means clustering algorithm to produce pseudo temporal annotations for anomalous videos and trained 3D ConvNet backbone for discriminative representation learning of anomaly instances.…”

Section: Related Workmentioning

confidence: 99%

“…Since training of GCN is computationally complex and can lead to unconstrained latent space, authors in [14] use clustering algorithms for cleaning the noisy labels of untrimmed anomaly videos. Different from [16], Zaheer et al [14] uses the k-means clustering algorithm to produce pseudo temporal annotations for anomalous videos and trained 3D ConvNet backbone for discriminative representation learning of anomaly instances. Since these methods are heavily dependent upon pseudo temporal annotations stage, so a noisy generation of temporal annotations can drastically mislead the training of 3D ConvNet backbone.…”

Section: Related Workmentioning

confidence: 99%

“…To combat this, recent anomaly detection approaches [11,17,15,7,4,16,13,14] adopt a weakly-supervised binary classification paradigm where both normal and anomaly videos are taken into account during training. In this setting, for a long untrimmed video sequence, only coarse video-level labels (i.e.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

DAM: Dissimilarity Attention Module for Weakly-supervised Video Anomaly Detection

Majhi

Das

Brémond

2021

2021 17th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)

View full text Add to dashboard Cite

Video anomaly detection under weak supervision is complicated due to the difficulties in identifying the anomaly and normal instances during training, hence, resulting in non-optimal margin of separation. In this paper, we propose a framework consisting of Dissimilarity Attention Module (DAM) to discriminate the anomaly instances from normal ones both at feature level and score level. In order to decide instances to be normal or anomaly, DAM takes local spatio-temporal (i.e. clips within a video) dissimilarities into account rather than the global temporal context of a video. This allows the framework to detect anomalies in real-time (i.e. online) scenarios without the need of extra window buffer time. Further more, we adopt two-variants of DAM for learning the dissimilarities between successive video clips. The proposed framework along with DAM is validated on two large scale anomaly detection datasets i.e. UCF-Crime and ShanghaiTech, outperforming the online state-of-the-art approaches by 1.5% and 3.4% respectively.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

DAM: Dissimilarity Attention Module for Weakly-supervised Video Anomaly Detection

Majhi

Das

Brémond

2021

2021 17th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)

View full text Add to dashboard Cite

show abstract

“…However, the input order of the features affects the convolution operation, which is not consistent with the characteristics of the gas path. A weakly supervised self-reasoning framework [ 13 ] has been proposed to improve the accuracy of video anomaly detection. It generates pseudo-labels by using binary clustering of spatiotemporal video features which helps in mitigating the noise present in the labels of anomalous videos.…”

Section: Introductionmentioning

confidence: 99%

A Weakly Supervised Gas-Path Anomaly Detection Method for Civil Aero-Engines Based on Mapping Relationship Mining of Gas-Path Parameters and Improved Density Peak Clustering

Sun

Zhong

2021

Sensors

View full text Add to dashboard Cite

Gas-path anomalies account for more than 90% of all civil aero-engine anomalies. It is essential to develop accurate gas-path anomaly detection methods. Therefore, a weakly supervised gas-path anomaly detection method for civil aero-engines based on mapping relationship mining of gas-path parameters and improved density peak clustering is proposed. First, the encoder-decoder, composed of an attention mechanism and a long short-term memory neural network, is used to construct the mapping relationship mining model among gas-path parameters. The predicted values of gas-path parameters under the restriction of mapping relationships are obtained. The deviation degree from the original values to the predicted values is regarded as the feature. To force the extracted features to better reflect the anomalies and make full use of weakly supervised labels, a weakly supervised cross-entropy loss function under extreme class imbalance is deployed. This loss function can be combined with a simple classifier to significantly improve the feature extraction results, in which anomaly samples are more different from normal samples and do not reduce the mining precision. Finally, an anomaly detection method is deployed based on improved density peak clustering and a weakly supervised clustering parameter adjustment strategy. In the improved density peak clustering method, the local density is enhanced by K-nearest neighbors, and the clustering effect is improved by a new outlier threshold determination method and a new outlier treatment method. Through these settings, the accuracy of dividing outliers and clustering can be improved, and the influence of outliers on the clustering process reduced. By introducing weakly supervised label information and automatically iterating according to clustering and anomaly detection results to update the hyperparameter settings, a weakly supervised anomaly detection method without complex parameter adjustment processes can be implemented. The experimental results demonstrate the superiority of the proposed method.

show abstract

Crime activities prediction system in video surveillance by an optimized deep learning framework

Kumar¹,

Reddy²

2022

Concurrency and Computation

View full text Add to dashboard Cite

The video surveillance technology is used to find crime events in public places and capture live public events. Hence, detecting the criminalist before the crime actions is the most needed event to catch the criminalist. However, the presence of noise content in the trained video has raised the difficulties in crime specification by maximizing the complexity range of the data. To overcome this issue, this research has designed a novel lion-based deep belief neural paradigm (LbDBNP) to identify criminals by their activities and handling tools. Initially, three types of datasets were trained to the system then the training flaws were eliminated in the preprocessing layer. Hereafter, the cleaned data is imported to the classification module to detect the crime events present in the video. Subsequently, the designed model is implemented using the Python framework in Windows 10 platform. To evaluate the efficiency of the designed model, the attack is launched in the proposed model after those metrics are calculated. In addition, the robustness of the designed system is verified by three datasets, such as UCSDped1, UCSDped2, and avenue crime. Also, the key parameters of the designed model have been evaluated and compared with other existing schemes to verify the proposed model's robustness by achieving the finest outcomes.

show abstract

A Self-Reasoning Framework for Anomaly Detection Using Video-Level Labels

Cited by 72 publications

References 27 publications

DAM: Dissimilarity Attention Module for Weakly-supervised Video Anomaly Detection

DAM: Dissimilarity Attention Module for Weakly-supervised Video Anomaly Detection

A Weakly Supervised Gas-Path Anomaly Detection Method for Civil Aero-Engines Based on Mapping Relationship Mining of Gas-Path Parameters and Improved Density Peak Clustering

Crime activities prediction system in video surveillance by an optimized deep learning framework

Contact Info

Product

Resources

About