Detecting Falls Inside the Building by Deep Learning
Article Sidebar
Main Article Content
Abstract
This article presents the detection of falls inside the building using deep learning. The method is to apply Convolutional Neural Network (CNN) for regularization by Dropout technique to create classification models. The data are collected in the form of video files from lnVia laboratory room, University de Franche-Comté, and separated into 60% of Training set, 20% of Test set, and 20% of Validation set. The three experiments are designed for different patterns according to the different inputs: 1) Grayscale images 2) Motion History Images (MHI), and 3) MHI combined with Grayscale image. Regarding the efficiency evaluation of the three best models, the classification model for Grayscale images has 0.9204 accuracy. The model for classifying MHI achieves 0.9193 accuracy. On the other hand, the model for classifying MHI combined with Grayscale images results in 0.9575 accuracy, which is the best model of all 3 experiments.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
References
Ahad, M. A. R. (2013). Motion history images for action recognition and understanding. Springer. https://doi.org/10.1007/978-1-4471-4730-5
Bangkok Hospital. (2021). Reducing fall-related accidents in the elderly. https://www.bangkokhospital.com/content/reduce-accidents-in-the-elderly-2
Eppel, S. (2017). Setting an attention region for convolutional neural networks using region selective features, for recognition of materials within glass vessels. arXiv. https://doi.org/10.48550/arXiv.1708.08711
Hu, M., Wang, H., Wang, X., Yang, J., & Wang, R. (2019). Video facial emotion recognition based on local enhanced motion history image and CNN-CTS-LSTM networks. Journal of Visual Communication and Image Representation, 62, 176-185.
Hwang, S., Ahn, D., Park, H., & Park, T. (2017). Poster abstract: Maximizing accuracy of fall detection and alert systems based on 3D convolutional neural network. Proceedings of the IEEE/ACM Second International Conference on Internet-of-Things Design and Implementation (IoTDI) (pp. 343-344). IEEE.
Jaroonphan, P., & Covavisaruch, N. (2013). Utilizing MHI for human’s gesture with repeating-path trajectory. Information Technology Journal, 9(2), 56-61. (in Thai)
Laboratoire ImViA. (2020). Fall detection dataset. https://imvia.u-bourgogne.fr/en/database/fall-detection-dataset-2.html
Li, X., Pang, T., Liu, W., & Wang, T. (2017). Fall detection for elderly person care using convolutional neural networks. Proceedings of the 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI) (pp. 1-6). IEEE.
Lu, N., Wu, Y., Feng, L., & Song, J. (2019). Deep learning for fall detection: Three-dimensional CNN combined with LSTM on video kinematic data. IEEE Journal of Biomedical and Health Informatics, 23(1), 314-323. https://doi.org/10.1109/JBHI.2018.2808281
Menacho, C., & Ordoñez, J. (2020). Fall detection based on CNN models implemented on a mobile robot. Proceedings of the 2020 17th International Conference on Ubiquitous Robots (UR) (pp. 284-289). IEEE.
Meng, H., Pears, N., Freeman, M., & Bailey, C. (2009). Motion history histograms for human action recognition. In B. Kisačanin, S. S. Bhattacharyya, & S. Chai (Eds.), Embedded Computer Vision (pp. 139-162). Springer. https://doi.org/10.1007/978-1-84800-304-0_7
Núñez-Marcos, A., Azkune, G., & Arganda-Carreras, I. (2017). Vision-based fall detection with convolutional neural networks. Wireless Communications and Mobile Computing, 2017, 3-17. https://doi.org/10.1155/2017/9474806.
Peng, D., Zhang, Y., & Guan, H. (2019). End-to-end change detection for high-resolution satellite images using improved UNet++. Remote Sensing, 11(11), Article 1382. https://doi.org/10.3390/rs11111382
Promrit, N., & Waijanya, S. (2021). Fundamental of deep learning in practice (1st ed.). Infopress Publisher, Bangkok. (In Thai)
Song, J., Zheng, J., Li, P., Lu, X., Zhu, G., & Shen, P. (2021). An effective multimodal image fusion method using MRI and PET for Alzheimer's disease diagnosis. Frontiers in Digital Health, 3, Article 637386. https://doi.org/10.3389/fdgth.2021.637386.
Subramanyan, T., Jang, Y., Tsogtbaatar, E., & Cho, S. (2020). Fall detection system for elderly people using vision-based analysis. Romanian Journal of Information Science and Technology, 23(1), 69-83.
Wang, K., Cao, G., Meng, D., Chen, W., & Cao, W. (2016). Automatic fall detection of human in video using combination of features. Proceedings of the IEEE International Conference on Bioinformatics and Biomedicine (BIBM) (pp. 1228-1233). IEEE.
Warunsiri, W., & Toontharm, K. (2017). Procentage rice inspection program based on image processing principles. Proceedings of the 10th Rajamangala University of Technology Tawan-ok Research Conference (pp. 56-62). Rajamangala University of Technology Tawan-ok, Chonburi, Thailand. (In Thai)
Zhang, Y., Dong, Z., Chen, X., Jia, W., Du, S., Muhammad, K., & Wang, S.-H. (2019). Image based fruit category classification by 13-layer deep convolutional neural network and data augmentation. Multimedia Tools and Applications, 78, 3613-3632.
Zou, S., Min, W., Liu, L., Wang, Q., & Zhou, X. (2017). Movement tube detection network integrating 3D CNN and object detection framework to detect fall. Electronics, 10(8), 2-16.
