Journal of the Korean Society of Industry Convergence (한국산업융합학회 논문집)

The Korean Society of Industry Convergence (한국산업융합학회)
권호 표지
DOI QR코드

DOI QR Code

YOLOv5 based Anomaly Detection for Subway Safety Management Using Dilated Convolution

  • Nusrat Jahan Tahira (Dept. of Electrical, Electronics and Communications Engineering, Kyungsung University) ;
  • Ju-Ryong Park (Dept. of Electrical, Electronics and Communications Engineering, Kyungsung University) ;
  • Seung-Jin Lim (Busan Transportaion Corporation) ;
  • Jang-Sik Park (Dept. of Electronic Engineering, Kyungsung University)
  • Received : 2022.12.07
  • Accepted : 2022.12.30
  • Published : 2023.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

With the rapid advancement of technologies, need for different research fields where this technology can be used is also increasing. One of the most researched topic in computer vision is object detection, which has widely been implemented in various fields which include healthcare, video surveillance and education. The main goal of object detection is to identify and categorize all the objects in a target environment. Specifically, methods of object detection consist of a variety of significant techniq ues, such as image processing and patterns recognition. Anomaly detection is a part of object detection, anomalies can be found various scenarios for example crowded places such as subway stations. An abnormal event can be assumed as a variation from the conventional scene. Since the abnormal event does not occur frequently, the distribution of normal and abnormal events is thoroughly imbalanced. In terms of public safety, abnormal events should be avoided and therefore immediate action need to be taken. When abnormal events occur in certain places, real time detection is required to prevent and protect the safety of the people. To solve the above problems, we propose a modified YOLOv5 object detection algorithm by implementing dilated convolutional layers which achieved 97% mAP50 compared to other five different models of YOLOv5. In addition to this, we also created a simple mobile application to avail the abnormal event detection on mobile phones.

Keywords

Acknowledgement

This work was supported by the BB21+ funded by Busan Metropolitan City and Busan Institute for Talent & Lifelong Education(BIT) and supported by "Human Activity Data of Unmaned Store" of AI learning data construction project through NIA(National Information Society Agency)

References

  1. M Paul, M. E. Shah and S. Chakraborty, "Human detection in surveillance videos and its applications-a review", EURASIP Journal on Advances in Signal Processing, pp. 176, 2013.
  2. Lo, Win-Tsung, Yue Shan Chang, Ruey-Kai Sheu, Chao-Tung Yang, Tong-Ying Juang, and Yu-Sheng Wu. "Implementation and evaluation of large-scale video surveillance system based on P2P architecture and cloud computing." International Journal of Distributed Sensor Networks 10, no. 4, 2014
  3. Litjens, G., et al.: Deep learning as a tool for increased accuracy and efficiency of histopathological diagnosis. Sci. Reports 6 (2016)
  4. Thomas Defard, Aleksandr Setkov, Angelique Loesch, Romaric Audigier, "PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2020.
  5. F. Mercaldo, F. Martinelli, and A. Santone, "A proposal to ensure social distancing with deep learning-based object detection," in 2021 International Joint Conference on Neural Networks (IJCNN). IEEE, 2021, pp. 1-5. 1-5.
  6. Krizhevsky A., Sutskever I., Hinton G.E., Imagenet classification with deep convolutional neural networks, Advances in Neural Information Processing Systems (2012), pp. 1097-1105.
  7. Zhang Q., Yang L.T., Chen Z., Li P. , A surv ey on deep learning for big data, Inf. Fusion, 42 (2018), pp. 146-157. https://doi.org/10.1016/j.inffus.2017.10.006
  8. Yuan J., Hou X., Xiao Y., Cao D., Guan W., Nie L., Multi-criteria active deep learning for image classification, Knowl.-Based Syst., 172 (2019), pp. 86-94 https://doi.org/10.1016/j.knosys.2019.02.013
  9. Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi, "You Only Look Once: Unified, Real-Time Object Detection., published at Computer Vision and Patter Recognition (CVPR)., 2015.
  10. Fisher Yu, Vladlen Koltun. "Multi-Scale Context Aggregation by Dilated Convolutions" Computer Vision and Pattern Recognition; 2016.
  11. Yulia Gavrilova, 'Convolutional Neural Networks for Beginners', Serokell, August 3rd, 2021, https://serokell.io/blog/introduction-to-convolutional-neural-networks
  12. 'Dilated Convolution', Geeksforgeeks, 02 Mar, 2022, https://www.geeksforgeeks.org/dilated-convolution/
  13. Benyamin Ghojogh, Maria N. Samad, Sayema Asif Mashhadi, Tania Kapoor, Wahab Ali, Fakhri Karray, Mark Crowley, "Feature Selection and Feature Extraction in Pattern Analysis: A Literature Review", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2019.
  14. Alexy Bochkovskiy, Chien-Yao Wang, Hong-Yuan Mark Liao; "YOLOv 4: Optimal Speed and Accuracy of object Detection". Computer Vision and Pattern Recognition(CVPR); 23 April 2020.
  15. Liu S.,. Qi L., Qin H., Shi J., Jia J. Path aggregation network for instance segmentation; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR); Salt Lake City, UT, USA. 18-23 June 2018; pp. 8759-8768.
  16. Shiv Ram Dubey, Satish Kumar Singh, Bidyut Baran Chaudhuri, "Activation Functions in Deep Learning: A Comprehensive Survey and Benchmark", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2021.
  17. Zhilu Zhang, Mert R. Sabuncu, "Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2018.