Name
Title | ||
|---|---|---|
Deep-Transfer-Learning-Based Abnormal Behavior Recognition Using Internet of Drones for Crowded Scenes |
Abstract | ||
|---|---|---|
Intelligent identification of abnormal behaviors in crowd scenes enables far more efficient development of smart cities. In recent approaches, abnormalities are detected using an autonomous unmanned aerial vehicle (UAV) by observing activity and representing crowd features such as density, direction, and anomalous behavior from acquired video frames. Thus, integrating fast monitoring via an autonomous UAV with automated procedures would significantly improve the effectiveness of detecting abnormal activities. Our work presents a real-time surveillance strategy for detecting anomalous behavior in crowds that is based on deep transfer learning and the Internet of Drones,and can be used to distinguish dynamic crowds. The deep structure is a simplified and lightweight representation of the ResNet learning structure. The proposed model that detects the abnormal behavior in this study processes the data analyzed from UAVs and intendS to increase its efficiency, robust-ness, and accuracy. For several UAV videos, the average accuracy is more than 90 percent. The experimental results reveal that the suggested method robustly detects abnormal behaviors in crowd scenarios using frames from UAV recordings with demanding conditions. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1109/IOTM.001.2100138 | IEEE Internet of Things Magazine |
DocType | Volume | Issue |
Journal | 5 | 2 |
ISSN | Citations | PageRank |
2576-3180 | 0 | 0.34 |
References | Authors | |
5 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Khosro Rezaee | 1 | 0 | 0.34 |
| Mohammad R. Khosravi | 2 | 26 | 7.55 |
| Maryam Saberi Anari | 3 | 0 | 0.34 |