Tools
Tools
Singh, Karan veer (2022) Identifying the Social distancing from Video Surveillance Cameras using Deep Learning Architectures. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (3MB) | Preview |
Preview |
PDF (Configuration manual)
Download (4MB) | Preview |
Abstract
In order to break the chain of spreading the COVID-19, social distancing is considered as an one of the effective measure. In order to encourage the people for practising the social distancing, we have proposed a social distancing monitoring framework which can identify the number of people violating the social distancing norms from surveillance camera videos in the real time. Our proposed framework trains the 3 different deep learning architectures over Open Image dataset. The architectures used for this research are Faster-RCNN, SSD-ResNet50 and YOLOv3. After performing the certain set of experiments, we have identified the YOLOv3 as the optimal model for object detection paradigm in order to identify the human in video sequences. The architecture identifies the people by adding bounding box information around the people. Using euclidean distance, the pairwise distance between the bounding boxes of their centroid is determined. The states are determined as safe, unsafe and moderated based on the closeness ratio between the individual.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning R Medicine > RA Public aspects of medicine > Public Health System |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Tamara Malone |
| Date Deposited: | 11 Mar 2023 13:45 |
| Last Modified: | 11 Mar 2023 13:45 |
| URI: | https://norma.ncirl.ie/id/eprint/6312 |
Actions (login required)
 |
View Item |