International Journal of Internet, Broadcasting and Communication

The Institute of Internet, Broadcasting and Communication (한국인터넷방송통신학회)
권호 표지
DOI QR코드

DOI QR Code

Detecting Abnormal Human Movements Based on Variational Autoencoder

  • Doi Thi Lan (Department of Electrical, Electronic and Computer Engineering, University of Ulsan) ;
  • Seokhoon Yoon (Department of Electrical, Electronic and Computer Engineering, University of Ulsan)
  • Received : 2023.06.10
  • Accepted : 2023.06.18
  • Published : 2023.08.31
Download PDF
⟨ Previous Next ⟩

Abstract

Anomaly detection in human movements can improve safety in indoor workplaces. In this paper, we design a framework for detecting anomalous trajectories of humans in indoor spaces based on a variational autoencoder (VAE) with Bi-LSTM layers. First, the VAE is trained to capture the latent representation of normal trajectories. Then the abnormality of a new trajectory is checked using the trained VAE. In this step, the anomaly score of the trajectory is determined using the trajectory reconstruction error through the VAE. If the anomaly score exceeds a threshold, the trajectory is detected as an anomaly. To select the anomaly threshold, a new metric called D-score is proposed, which measures the difference between recall and precision. The anomaly threshold is selected according to the minimum value of the D-score on the validation set. The MIT Badge dataset, which is a real trajectory dataset of workers in indoor space, is used to evaluate the proposed framework. The experiment results show that our framework effectively identifies abnormal trajectories with 81.22% in terms of the F1-score.

Keywords

Acknowledgement

This work was supported by the Institute of Information and Communication Technology Planning and Evaluation (IITP) Grant by the Korean Government through MSIT (Development of 5G-Based Shipbuilding and Marine Smart Communication Platform and Convergence Service) under Grant 2020-0-00869.

References

  1. Matousek M, Mohamed EZ, Kargl F, Bosch C. " Detecting anomalous driving behavior using neural networks. " In 2019 IEEE Intelligent Vehicles Symposium (IV) 2019 Jun 9 (pp. 2229-2235). IEEE. DOI: 10.1109/IVS.2019.8814246.
  2. Zhang M, Chen C, Wo T, Xie T, Bhuiyan MZ, Lin X. " SafeDrive: Online driving anomaly detection from largescale vehicle data. " IEEE Transactions on Industrial Informatics. 2017 Feb 24;13(4):2087-96. DOI: 10.1109/TII.2017.2674661.
  3. Dai J, Yang B, Guo C, Ding Z. " Personalized route recommendation using big trajectory data." In 2015 IEEE 31st international conference on data engineering 2015 Apr 13 (pp. 543-554). IEEE. DOI:10.1109/ICDE.2015.7113313
  4. Karatas GB, Karagoz P, Ayran O. "Trajectory pattern extraction and anomaly detection for maritime vessels." Internet of Things. 2021 Dec 1; 16:100436. DOI: doi.org/10.1016/j.iot.2021.100436
  5. Lee JG, Han J, Li X. " Trajectory outlier detection: A partition-and-detect framework." In 2008 IEEE 24th International Conference on Data Engineering 2008 Apr 7 (pp. 140-149). IEEE. DOI: 10.1109/ICDE.2008.4497422.
  6. Zhu, Z.; Yao, D.; Huang, J.; Li, H.; Bi, J. " Sub-trajectory-and trajectory-neighbor-based outlier detection over trajectory streams." In Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining, Melbourne, VIC, Australia, 3-6 June 2018. DOI: 10.1109/CGiV.2016.65
  7. Wang, Yulong, et al. "Detecting anomalous trajectories and behavior patterns using hierarchical clustering from taxi GPS data." ISPRS International Journal of Geo-Information 7.1 (2018): 25. DOI: doi.org/10.3390/ijgi7010025
  8. Song L, Wang R, Xiao D, Han X, Cai Y, Shi C. Anomalous trajectory detection using recurrent neural network. In Advanced Data Mining and Applications: 14th International Conference, ADMA 2018, Nanjing, China, November 16-18, 2018, Proceedings 14 2018 (pp. 263-277). Springer International Publishing.
  9. Cheng Y, Wu B, Song L, Shi C. Spatial-temporal recurrent neural network for anomalous trajectories detection. In Advanced Data Mining and Applications: 15th International Conference, ADMA 2019, Dalian, China, November 21-23, 2019, Proceedings 15 2019 (pp. 565-578). Springer International Publishing.
  10. Kieu T, Yang B, Jensen CS. Outlier detection for multidimensional time series using deep neural networks. In2018 19th IEEE international conference on mobile data management (MDM) 2018 Jun 25 (pp. 125-134). IEEE. DOI: 10.1109/MDM.2018.00029.
  11. M. Schuster and K. K. Paliwal, "Bidirectional recurrent neural networks," IEEE Transactions on Signal Processing, vol. 45, no. 11, pp. 2673-2681, 1997. DOI: 10.1109/78.650093
  12. Matias P, Folgado D, Gamboa H, Carreiro AV. Robust Anomaly Detection in Time Series through Variational AutoEncoders and a Local Similarity Score. In Biosignals 2021 (pp. 91-102). DOI: 10.5220/0010320500002865
  13. Tao Y, Both A, Silveira RI, Buchin K, Sijben S, Purves RS, Laube P, Peng D, Toohey K, Duckham M. A comparative analysis of trajectory similarity measures. GIScience & Remote Sensing. 2021 Jul 4;58(5):643-69. DOI: doi.org/10.1080/15481603.2021.1908927
  14. Olguin DO, Waber BN, Kim T, Mohan A, Ara K, Pentland A. "Sensible organizations: Technology and methodology for automatically measuring organizational behavior." IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics) 39.1 (2008): 43-55. DOI: 10.1109/TSMCB.2008.2006638
  15. SzekeR, M. "Spatio-Temporal Outlier Detection in Streaming Trajectory Data." Master's Thesis, School of Computer Science and Communication (CSC), Stockholm, Sweeden, 2014.
  16. Banerjee, P.; Talwalkar, P.; Ranu, S. "Mantra: A scalable approach to mining temporally anomalous subtrajectories." In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13-17 August 2016; pp. 1415-1424. DOI: 10.1145/2939672.2939846
  17. Pereira J, Silveira M. " Learning representations from healthcare time series data for unsupervised anomaly detection." In 2019 IEEE international conference on big data and smart computing (BigComp) 2019 Feb 27 (pp. 1-7). IEEE. DOI: 10.1109/BIGCOMP.2019.8679157