本研究中提出使用較有隱私的低解析度熱感測器，藉由深度學習模型以辨識低解析度熱圖像，即時的檢測跌倒。本研究中使用32x24熱電堆紅外線熱感測器，可顯示人體輪廓。經過影像預處理後輸出數據，搭配神經網路模型進行圖像辨識分類。模型中使用許多架構組合，包含殘差神經網路(Residual neural network)、雙向長短期記憶網路(Bidirectional Long Short-Term Memory Networks)和注意力機制模型(Attention)，透過多層網路架構對數據進行評估。本論文使用公開的資料集(eHomeSeniors Dataset)訓練模型，此資料集包含15種跌倒姿態，經過整理後取得442個正常姿態影像與399個跌倒姿態影像。以此資料集70%訓練模型A (ResNet18-BiLSTM-Attention)和模型B (ResNet18-LSTM)，15%為驗證集，最後15%為測試集以測試模型的準確率。驗證方式使用混淆矩陣分析，HOG-SVM 為97.8% ，KNN為98%，模型A與模型B準確度都獲得99.2%，再比對F-score，模型A為0.991，模型B為0.99，總體分數上模型A優於模型B。模型A在實地場景實驗中也獲得97.5%的準確性。

In this research, a low-resolution thermopile array sensor is used to take pictures., And the deep learning model is used to recognize the images, and to perform fall detection. The 32x24 thermopile array sensor is used to display the contour of the body. After image preprocessing, the neural network models are used for image recognition and classification. Residual neural network, Bidirectional Long Short-Term Memory Networks and Attention are used in the proposed model. The open dataset (eHomeSeniors Dataset) was used to train the model. The dataset contains 15 kinds of fall postures. After monitored classification, 442 normal posture images and 399 fall posture images were obtained. Proposed model A (ResNet18-BiLSTM-Attention) and proposed model B (ResNet18-LSTM) were trained by 70% of the dataset; 15% of the dataset is used for validation, and the last 15% is the test set to test the accuracy of the model. Confusion matrix was used to verify accuracy: HOG-SVM is 97.8%, KNN is 98%, model A and model B both get 99.2% accuracy. As for the F-Score, model A is 0.991, and model B is 0.99. Overall model A is better than model B. Model A also achieved an accuracy of 97.5% in field experiments.

[1] "World Population Ageing 2020 Highlights," United Nations, 2020. [Online]. Available: https://www.un.org/development/desa/pd/sites/www.un.org.development.desa.pd/files/files/documents/2020/Sep/un_pop_2020_pf_ageing_10_key_messages.pdf, (last retrieved:October, 2020).
[2] "中華民國人口推估（2020至2070年）," 中華民國國家發展委員會, [Online]. Available: https://pop-proj.ndc.gov.tw/upload/download/中華民國人口推估(2020至2070年)報告.pdf, (last retrieved: August. 4, 2020).
[3] 統計處, "108年死因統計結果分析," 中華民國衛生福利部, 109. [Online]. Available: https://www.mohw.gov.tw/dl-61889-aeff6df1-4a21-46ce-a37c-19600428cd35.html, (last retrieved: June. 16, 2020).
[4] Y.-C. Chen and Y.-W. Lin, "Indoor RFID gait monitoring system for fall detection," 2010 2nd International Symposium on Aware Computing, pp. 207-212, 2010.
[5] J. Chen, K. Kwong, D. Chang, J. Luk and R. Bajcsy, "Wearable sensors for reliable fall detection," 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference, pp. 3551-3554, 2006.
[6] M. Zhang and A. A. Sawchuk, "Human daily activity recognition with sparse representation using wearable sensors," IEEE journal of Biomedical and Health Informatics, vol. 17, no. 3, pp. 553-560, 2013.
[7] A. M. Khan, Y.-K. Lee, S. Y. Lee and T.-S. Kim, "A triaxial accelerometer-based physical-activity recognition via augmented-signal features and a hierarchical recognizer," IEEE transactions on information technology in biomedicine, vol. 14, no. 5, pp. 1166-1172, 2010.
[8] Y. Zigel, D. Litvak and I. Gannot, "A method for automatic fall detection of elderly people using floor vibrations and sound—Proof of concept on human mimicking doll falls," IEEE transactions on biomedical engineering, vol. 56, no. 12, pp. 2858-2867, 2009.
[9] J. Shim, M.-h. Shim, Y.-s. Baek and T.-d. Han, "The development of a detection system for seniors' accidental fall from bed using cameras," Proceedings of the 5th International Conference on Ubiquitous Information Management and Communication, pp. 1-4, 2011.
[10] C. Rougier, J. Meunier, A. St-Arnaud and J. Rousseau, "Fall detection from human shape and motion history using video surveillance," 21st International Conference on Advanced Information Networking and Applications Workshops (AINAW'07), pp. 875-880, 2007.
[11] G. Diraco, A. Leone and P. Siciliano, "An active vision system for fall detection and posture recognition in elderly healthcare," 2010 Design, Automation & Test in Europe Conference & Exhibition (DATE 2010), pp. 1536-1541, 2010.
[12] T. Liu, X. Guo and G. Wang, "Elderly-falling detection using distributed direction-sensitive pyroelectric infrared sensor arrays," Multidimensional Systems and Signal Processing, vol. 23, no. 4, pp. 451-467, 2012.
[13] W.-H. Chen and H.-P. Ma, "A fall detection system based on infrared array sensors with tracking capability for the elderly at home," 2015 17th International Conference on E-health Networking, Application & Services (HealthCom), pp. 428-434, 2015.
[14] X. Fan, H. Zhang, C. Leung and Z. Shen, "Robust unobtrusive fall detection using infrared array sensors," 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, pp. 194-199, 2017.
[15] N. Lu, Y. Wu, L. Feng and J. Song, "Deep learning for fall detection: three-dimensional CNN combined with LSTM on video kinematic data," IEEE Journal of Biomedical and Health Informatics, vol. 23, no. 1, pp. 314-323, 2018.
[16] A. Abobakr, M. Hossny, H. Abdelkader and S. Nahavandi, "RGB-D fall detection via deep residual convolutional LSTM networks," 2018 Digital Image Computing: Techniques and Applications (DICTA), pp. 1-7, 2018.
[17] M. Gochoo, T.-H. Tan, S.-H. Liu, F.-R. Jean, F. S. Alnajjar and S.-C. Huang, "Unobtrusive activity recognition of elderly people living alone using anonymous binary sensors and DCNN," IEEE journal of biomedical and health informatics, vol. 23, no. 2, pp. 693-702, 2018.
[18] T.-H. Tan, M. Gochoo, S.-C. Huang, Y.-H. Liu, S.-H. Liu and Y.-F. Huang, “Multi-resident activity recognition in a smart home using RGB activity image and DCNN,” IEEE Sensors Journal, vol. 18, no 23, pp. 9718-9727, 2018.
[19] M. Gochoo, T.-H. Tan, S.-C. Huang, T. Batjargal, J.-W. Hsieh, F. S. Alnajjar and Y.-F. Chen, "Novel IoT-based privacy-preserving yoga posture recognition system using low-resolution infrared sensors and deep learning," IEEE Internet of Things Journal, vol. 6, no. 4, pp. 7192-7200, 2019.
[20] M. Gochoo, T.-H. Tan, F. Alnajjar, J.-W. Hsieh and P.-Y. Chen, "Lownet: Privacy preserved ultra-low resolution posture image classification," 2020 IEEE International Conference on Image Processing (ICIP), pp. 663-667, 2020.
[21] M. M. Islam, O. Tayan, M. R. Islam, M. S. Islam, S. Nooruddin, M. N. Kabir and M. R. Islam, "Deep learning based systems developed for fall detection: a review," IEEE Access, vol. 8, pp. 166117-166137, 2020.
[22] B. Kwolek and M. Kepski, "Human fall detection on embedded platform using depth maps and wireless accelerometer," Computer methods and programs in biomedicine, vol. 117, no. 3, pp. 489-501, 2014.
[23] F. Riquelme, C. Espinoza, T. Rodenas, J.-G. Minonzio and C. Taramasco, "eHomeSeniors Dataset: An Infrared thermal sensor dataset for automatic fall detection research," Sensors, vol. 19, no. 20, 2019.
[24] "Gaussian Smoothing," Archive, Internet, [Online]. Available: https://web.archive.org/web/20061109221710/http://www.cee.hw.ac.uk/hipr/html/gsmooth.html, (last retrieved: November. 9, 2006).
[25] M. Minsky and S. Papert, Perceptrons, Cambridge: M.I.T. Press, 1969.
[26] Y. LeCun, L. Bottou, Y. Bengio and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[27] D. Scherer, A. Müller and S. Behnke, "Evaluation of pooling operations in convolutional architectures for object recognition," International conference on artificial neural networks, pp. 92-101, 2010.
[28] V. Nair and G. E. Hinton, "Rectified linear units improve restricted boltzmann machines," Proceedings of the 27th international conference on machine learning (ICML-10), 2010.
[29] O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla and M. Bernstein, "Imagenet large scale visual recognition challenge," International journal of computer vision, vol. 115, no. 3, pp. 211-252, 2015.
[30] K. He, X. Zhang, S. Ren and J. Sun, "Deep residual learning for image recognition," Computer Vision and Pattern Recognition (CVPR), pp. 770-778, 2016.
[31] Y. Bengio, P. Simard and P. Frasconi, "Learning long-term dependencies with gradient descent is difficult," IEEE transactions on neural networks, vol. 5, no. 2, pp. 157-166, 1994.
[32] S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[33] A. Graves and J. Schmidhuber, "Framewise phoneme classification with bidirectional LSTM and other neural network architectures," Neural Networks, vol. 18, no. 5-6, pp. 602-610, 2005.
[34] P. Zhou, W. Shi, J. Tian, Z. Qi, B. Li, H. Hao and B. Xu, "Attention-based bidirectional long short-term memory networks for relation classification," Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pp. 207-212, 2016.
[35] Y. Xiao and X. Lin, "Fall detection using low-resolution thermal sensor," Department of Electrical and Computer Engineering, Boston University, Boston. [Online]. Available: http://www.bu.edu/vip/files/pubs/reports/YXXL19-03buece.pdf, (last retrieved: April. 10, 2019)
[36] A. Krizhevsky, I. Sutskever and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," Communications of the ACM, vol. 60, no. 6, pp. 84-90, 2017.
[37] Melexis Ltd, "MLX90640 32×24 IR array Datasheet," 2019. [Online]. Available: https://www.melexis.com/-/media/files/documents/datasheets/mlx90640-datasheet-melexis.pdf, (last retrieved: December. 3, 2019)
[38] K. He and J. Sun, "Convolutional neural networks at constrained time cost," Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 5353-5360, 2015.
[39] C. Taramasco, T. Rodenas, F. Martinez, P. Fuentes, R. Munoz, R. Olivares, V. H. C. De Albuquerque and J. Demongeot, "A novel monitoring system for fall detection in older people," IEEE Access, vol. 6, pp. 43563-43574, 2018.
[40] M. Schuster and K. Paliwal, "Bidirectional recurrent neural networks," IEEE Transactions on Signal Processing, vol. 45, no. 11, pp. 2673-2681, 1997.
[41] 黃日鉦, 人工智慧與深度學習理論與Python實踐, 台北: 碁峯資訊股份有限公司, 2020. ISBN:9789865023898
[42] 唐進民, AI視覺大全：用最好用的PyTorch實作, 台北: 佳魁資訊股份有限公司, 2018, p. 296. ISBN:9789863796978