A Survey on Video Anomaly Detection in Surveillance System

Surveillance video anomaly detection is a challenging and important task that aims to identify unusual or suspicious events in video streams without human intervention. Anomaly detection can help enhance the security and safety of various critical infrastructure systems, such as football stadiums, a...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	IEEE Recent Advances in Intelligent Computational Systems s. 1 - 5
Hlavní autoři:	T, Manesh, Nataraj, Neetha K, Jayaraj, Akshaya, Joby, Alen, Ananthakrishnan, P V, Thankachan, Basil K
Médium:	Konferenční příspěvek
Jazyk:	angličtina
Vydáno:	IEEE 16.05.2024
Témata:	CNN Mask R-CNN One class detection R-CNN Safety Security Spatio Temporal autoencoders Streaming media Surveillance Surveys Training Weapons
ISSN:	2769-5565
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Surveillance video anomaly detection is a challenging and important task that aims to identify unusual or suspicious events in video streams without human intervention. Anomaly detection can help enhance the security and safety of various critical infrastructure systems, such as football stadiums, and cricket stadiums. However, anomaly detection also faces many difficulties, such as the diversity and complexity of normal and abnormal behaviors, the lack of sufficient and consistent annotations, and the high computational cost of processing large-scale video data. In this survey, we review the recent advances and challenges of artificial intelligence techniques in surveillance video anomaly detection. We highlight the strengths and limitations of different techniques, such as deep learning, generative models, graph neural networks, and multiple instance learning. We conclude by providing some future research directions and open issues in this field.
ISSN:	2769-5565
DOI:	10.1109/RAICS61201.2024.10690095