Comparative Study for Anomaly Detection in Crowded Scenes | ||||
| International Journal of Intelligent Computing and Information Sciences | ||||
| Article 15, Volume 21, Issue 3, November 2021, Page 84-94 PDF (587.36 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/ijicis.2021.84588.1112 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
Mohamed Abdelghafour   1; Maryam ElBery 2; Zaki Taha1
| ||||
| 1Computer Science department, Computer and Information Science, Ain Shams University, Cairo, Egypt | ||||
| 2Scientific Computing department, Computer and Information Science, Ain Shams University, Cairo, Egypt | ||||
| Abstract | ||||
| Nowadays, video analysis is an important research area especially from a security point of view. The discovery of unusual activities is important because it is a difficult task for humans especially with increasing number of surveillance cameras in all crowded places. That is because it requires a lot of human effort, and these activities happen rarely. Also the definition of anomaly events is different based on the location of the event. For example running in the park is a normal event but running in a restaurant is an abnormal event. The event is the same but the place was the factor of making it normal or not. The main objective of this paper is to compile what has been achieved in the field of anomaly detection and compare them, and to look at the different datasets used in the recent period. We will show how to detect and identify anomalies in videos, approaches for video anomaly detection and also what are the latest learning frameworks. | ||||
| Keywords | ||||
| Abnormal event detection; Video surveillance; Unsupervised learning | ||||
|
Statistics Article View: 266 PDF Download: 415 |
||||
