Multiple instance-based video anomaly detection using deep temporal encoding–decoding

Kamoona, Ammar Mansoor; Gostar, Amirali Khodadadian; Bab–Hadiashar, Alireza; Hoseinnezhad, Reza

doi:10.1016/j.eswa.2022.119079

Cited by 34 publications

(14 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Combine a non-linear SVM classifier and the representation power of I3D deep features to build a robust and competitive estimator. • Compare our results with two previous state-of-the-art studies Wan et al [2020] and Kamoona et al [2020] on the application of MIL paradigm and binary classification for video anomaly detection. From the evaluation and comparison of the proposed approach with two studies in the literature, we achieve promising results.…”

Section: Introductionsupporting

confidence: 57%

“…From the obtained results, the proposed approach overcomes other anomaly detection state-of-the-art approaches in performance. Kamoona et al (2020) suggested a deep neural network with an encoding-decoding architecture for anomaly detection in video surveillance scenarios to allow the capture of temporal and spatial information of video instances. The main contribution was to consider the temporal relations among video instances to treat them as sequential visual data instead of a set of independent instances.…”

Section: Related Workmentioning

confidence: 99%

“…The results of the last two research studies Wan et al [2020] and Kamoona et al [2020] were used to compare the achieved results in our proposed approach since both use the same benchmark dataset applied in our study. Similar to Wan et al [2020], our study also evaluates the efficiency of binary classification and MIL approach for video anomaly detection.…”

Section: Related Workmentioning

confidence: 99%

“…Anomaly detection is a crucial task in many domain applications, such as intrusion detection, frauds, and video surveillance Li et al [2021]. The manual processing of the information produced by surveillance cameras demands excessive manpower, which makes it relevant to automate video anomaly detection Kamoona et al [2020]. Due to the subjectivity in anomalies definition, it may be difficult to specify once it depends on the location and situation and varies widely in content and duration Wan et al [2020].…”

Section: Introductionmentioning

confidence: 99%

“…Deep learning approaches have achieved remarkable results in different application domains, such as image classification, object detection, voice processing, and anomaly detection Pawar and Attar [2019]. Unsupervised deep learning methods have been applied for model learning and feature extraction once distinct spatiotemporal handcrafted attributes (such as color, optical flow, and texture, commonly used in anomaly detection procedures) do not produce good generalization Kamoona et al [2020]. Researchers proposed many deep learning-based classification methods and feature extraction approaches in recent years Perera and Patel [2019].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Weakly Supervised Video Anomaly Detection Combining Deep Features with Shallow Neural Networks

Pereira

Maia

2022

JBCS

View full text Add to dashboard Cite

Deep features have outgrown hand-sketched features in many applications. The availability of pre-trained deep feature extractors helps to overcome one of the deep learning main drawbacks, which is the need for large volumes of data for training. Multiple Instance Learning (MIL) has become an attractive solution for video surveillance literature once it allows working with weakly labeled bases. This work evaluates a video anomaly detection approach based on the MIL paradigm combining deep features with shallow Neural Networks. For computational efficiency, we apply Principal Component Analysis (PCA) for dimensionality reduction before classification. We performed the experiments from a set of I3D (Inflated 3D) features, which corresponds to the ShanghaiTech benchmark dataset, and the MLP and SVM shallow classifiers achieved competitive results.

show abstract

Section: Introductionsupporting

confidence: 57%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Weakly Supervised Video Anomaly Detection Combining Deep Features with Shallow Neural Networks

Pereira

Maia

2022

JBCS

View full text Add to dashboard Cite

show abstract

Deep learning based anomaly detection in real-time video

Elmetwally,

Eldeeb,

Elmougy

2024

Multimed Tools Appl

View full text Add to dashboard Cite

Many security cameras have been put up in places like airports, roads, and banks for the safety of these public places. These cameras make a lot of video data, and most security camera recordings are only ever seen when something strange happens. This means that monitoring has to be done by people, which is time-consuming and often wrong, so automatic ways of monitoring have to be used. In this paper, we propose a system that automatically detects irregular events in videos based on the integration of Inflated 3D Convolution Network (I3D-ResNet50) and deep Multiple Instance Learning (MIL). This system considers both regular and unusual videos as negative and positive packets, respectively. Each video snippet is a case of that packet. An anomaly score is generated for each video snippet using a fully connected Neural Network (NN). After processing videos, we used an I3D-ResNet50 to extract features after applying 10-crop augmentations to the UCF-101 dataset that contains 130 GB of videos with 13 abnormal events such as fighting, stealing, abuse, etc., as well as normal events. Our experimental results show that the AUC is 82.85% with only 10,000 iterations compared with other approaches. This means that our model is better at spotting anomalies in real-time videos.

show abstract