臺灣博碩士論文加值系統

本文包含兩研究主題，第一個主題是基於長短期記憶網路(Long Short-Term Memory, LSTM)，有效地將連續的人體關鍵點分類成動作。在動作識別實驗的樣本以柔道的3個競技動作搭配6個一般動作來做動作識別，實際結果可得92%的準確率。第二個主題是基於K-近鄰演算法(K-Nearest Neighbor Classification, KNN)，KNN是人工智慧中的一種分類方法，透過特徵值比較快速的預測一個樣本的可能分類。在位置識別實驗的樣本以台北市為區域，以知名社群網路的地理位置點(geographical points)。實驗的結果顯示我們提出權重KNN方法準確率可達99.5%以上。

This paper contains two research topics. The first topic is to effectively classification continuous human body key points into actions based on Long Short-Term Memory (LSTM). In the action recognition experiment, the three competitive actions of Judo combined with six general actions are recognized. The actual result can achieve 92% accuracy. The second topic is based on K-Nearest Neighbor Classification (KNN). KNN is a classification method in artificial intelligence, which can quickly predict the possible classification of a sample through feature values. The sample in the location recognition experiment uses Taipei City as the area and uses geographic points of well-known social networks. The experimental results show that the accuracy of our proposed weighted KNN method can reach more than 99.5%.

目錄
摘要 I
致謝 III
目錄 IV
圖目錄 VII
表目錄 IX
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機與目的 3
1.3 論文章節介紹 5
第二章 相關研究 6
2.1 機器學習(Machine Learning) 6
2.1.1 K-近鄰演算法(K-Nearest Neighbor) 7
2.2 人工神經網路(Artificial Neural Network) 8
2.2.1 激勵函數 9
2.2.2 最佳化演算法 10
2.2.3 損失函數 11
2.3 卷積神經網路(Convolutional Neural Network) 12
2.3.1 卷積層(Convolutional Layer) 12
2.3.2 池化層(Pooling Layer) 13
2.4 遞歸神經網路(Recurrent Neural Network) 14
2.5 長短期記憶網路(Long Short-Term Memory Network) 15
2.6 評估指標 16
2.7 人體骨架提取(OpenPose) 18
2.8 定位技術 19
2.8.1 衛星導航系統(Global Navigation Satellite System) 20
2.8.2 全球定位系統(Global Positioning System) 20
2.9 文獻探討 22
2.9.1 動作識別 22
2.9.2 位置識別 25
第三章 識別方法 28
3.1 動作識別 28
3.1.1 資料預處理 30
3.1.2 模型架構 33
3.2 位置識別 35
3.2.1 位置識別 36
3.2.2 位置預測 38
第四章 實驗結果 41
4.1 動作識別實驗結果 41
4.1.1 資料集介紹 41
4.1.2 數據集預處理之比較 43
4.1.3 模型最佳參數設置 44
4.1.4 RNN與LSTM模型比較 50
4.1.5 LSTM模型混淆矩陣準確度 51
4.2 位置識別實驗結果 52
4.2.1 非加權KNN與加入權重KNN之比較 54
第五章 結論 59
參考文獻 60

圖目錄
圖2.1 KNN演算法 7
圖2.2 人工神經元模型 8
圖2.3 卷積神經網路(CNN) 12
圖2.4 卷積運算 13
圖2.5 Max Pooling 13
圖2.6 RNN網路 14
圖2.7 LSTM架構 15
圖2.8 混淆矩陣 17
圖3.1 動作識別流程 29
圖3.2 OpenPose絕對座標 31
圖3.3 OpenPose人體關鍵點 31
圖3.4 正規化人體關鍵點 32
圖3.5 LSTM網路架構 34
圖3.6 邊緣相交算法 37
圖3.7 多邊形內外算法 37
圖3.8 具有地理點的多邊形 39
圖4. 1預處理準確度之比較 43
圖4.2 訓練集資料量50筆準確度 44
圖4.3 訓練集資料量100筆準確度 45
圖4.4 訓練集資料量150筆準確度 46
圖4.5 訓練集資料量200筆準確度 46
圖4.6 訓練集資料量250筆準確度 47
圖4.7 訓練集資料量300筆準確度 48
圖4.8 各訓練集資料量驗證準確度 48
圖4.9 隱藏層節點數量設置 49
圖4.10 RNN、LSTM模型比較 50
圖4.11 混淆矩陣結果 51
圖4.12 地理點分布 53
圖4.13 K=3準確度比較 54
圖4.14 K=5準確度比較 55
圖4.15 K=7準確度比較 55
圖4.16 K=11準確度比較 56

表目錄
表3.1 模型參數 34
表4.1 動作資料集 42
表4.2 混淆矩陣結果 52
表4.3 加權和非加權預測方法 57

[1]P. Zhang, C. Lan, J. Xing, W. Zeng, J. Xue, and N. Zheng, "View Adaptive Recurrent Neural Networks for High Performance Human Action Recognition from Skeleton Data, "IEEE International Conference on Computer Vision, 2017, pp. 2117-2126.
[2]H. Zhang, Y. Song and Y. Zhang, "Graph Convolutional LSTM Model for Skeleton-Based Action Recognition," IEEE International Conference on Multimedia and Expo (ICME), Shanghai, China, 2019, pp. 412-417
[3]S. Song, C. Lan, J. Xing, W. Zeng and J. Liu, "Spatio-Temporal Attention-based LSTM Networks for 3D Action Recognition and Detection," in IEEE Transactions on Image Processing, Vol. 27, No. 7, pp. 3459-3471.
[4]S. Jun and Y. Choe, "Deep Batch-Normalized LSTM Networks with Auxiliary Classifier for Skeleton based Action Recognition,"IEEE International Conference on Image Processing, Applications and Systems (IPAS), Sophia Antipolis, France, 2018, pp. 279-284.
[5]X. Ouyang et al., "A 3D-CNN and LSTM Based Multi-Task Learning Architecture for Action Recognition," IEEE Access, Vol. 7, pp. 40757-40770.
[6]H. Pan and Y. Chen, "Multilevel LSTM for Action Recognition Based on Skeleton Sequence,"IEEE 21st International Conference on High Performance Computing and Communications; IEEE 17th International Conference on Smart City; IEEE 5th International Conference on Data Science and Systems (HPCC/SmartCity/DSS), Zhangjiajie, China, 2019, pp. 2218-2223.
[7]E. Martinez-Martin and M. Cazorla, "A Socially Assistive Robot for Elderly Exercise Promotion," in IEEE Access, Vol. 7, pp. 75515-75529.
[8]S. Yamanaka, C. Lee and S. Date, "A Parallel LSTM-Based Missing Body Feature Point Completion in Video Frames,"International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, USA, 2019, pp. 646-651.
[9]N. C. Tay, C. Tee, T. S. Ong and P. S. Teh, "Abnormal Behavior Recognition using CNN-LSTM with Attention Mechanism,"1st International Conference on Electrical, Control and Instrumentation Engineering (ICECIE), Kuala Lumpur, Malaysia, 2019, pp. 1-5.
[10]C. Chen, R. Jafari and N. Kehtarnavaz, "A Real-Time Human Action Recognition System Using Depth and Inertial Sensor Fusion," in IEEE Sensors Journal, Vol. 16, No. 3, pp. 773-781
[11]C. Wang, C. Chiang, J. Ding and J. Wang, "Dynamic Tracking Attention Model for Action Recognition,"IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, LA, 2017, pp. 1617-1621.
[12]C. Si, W. Chen, W. Wang, L. Wang and T. Tan, "An Attention Enhanced Graph Convolutional LSTM Network for Skeleton-Based Action Recognition,"IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 2019, pp. 1227-1236.
[13]Y. Han, S. Chung, A. Ambikapathi, J. Chan, W. Lin and S. Su, "Robust Human Action Recognition Using Global Spatial-Temporal Attention for Human Skeleton Data,"International Joint Conference on Neural Networks (IJCNN), Rio de Janeiro, 2018, pp. 1-8.
[14]N. Dawar and N. Kehtarnavaz, "Action Detection and Recognition in Continuous Action Streams by Deep Learning-Based Sensing Fusion," IEEE Sensors Journal, Vol. 18, No. 23, pp. 9660-9668.
[15]H. Wei and N. Kehtarnavaz, "Simultaneous Utilization of Inertial and Video Sensing for Action Detection and Recognition in Continuous Action Streams," IEEE Sensors Journal, Vol. 20, No. 11, pp. 6055-6063.
[16]H. Xu, C. Yuan, P. Li and Y. Wang, "Design and implementation of action recognition system based on RFID sensor,"13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD), Guilin, 2017, pp. 3021-3025.
[17]J. Li, X. Liu, J. Xiao, H. Li, S. Wang and L. Liu, "Dynamic Spatio-Temporal Feature Learning via Graph Convolution in 3D Convolutional Networks," International Conference on Data Mining Workshops (ICDMW), Beijing, China, 2019, pp. 646-652.
[18]X. Gao et al., "3D Skeleton-Based Video Action Recognition by Graph Convolution Network,"IEEE International Conference on Smart Internet of Things (SmartIoT), Tianjin, China, 2019, pp. 500-501.
[19]R. Liu, C. Xu, T. Zhang, W. Zhao, Z. Cui and J. Yang, "Si-GCN: Structure-induced Graph Convolution Network for Skeleton-based Action Recognition,"International Joint Conference on Neural Networks (IJCNN), Budapest, Hungary, 2019, pp. 1-8.
[20]R. Zhao, K. Wang, H. Su and Q. Ji, "Bayesian Graph Convolution LSTM for Skeleton Based Action Recognition,"IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Korea (South), 2019, pp. 6881-6891.
[21]J. Tang, Z. Liu, Y. Wang, J. Yang and Q. Wang, "Using Geographic Information and Point of Interest to Estimate Missing Second-Hand Housing Price of Residential Area in Urban Space. "IEEE International Smart Cities Conference (ISC2), 2018, pp. 1-8.
[22]M. I. Santoso, Z. Arham and E. Khudzaeva, "Design of a web-based Geographic Information Systems spatial for distribution of historic site. "4th International Conference on Cyber and IT Service Management, 2016, pp. 1-5.
[23]A. S. Dahish and A. Ahmad, "An application of Geographical Information System and Remote Sensing techniques for detection of oil spill. "17th International Conference on Geoinformatics, 2009, pp. 1-4.
[24]R. Ouaret, B. Birregah, E. Soulier, S. Auclair and F. Boulahya, " Random Forest location prediction from social networks during disaster events." Sixth International Conference on Social Networks Analysis, Management and Security (SNAMS), 2019, pp. 535-540.
[25]Y. Huang, C. Li and S. Jeng, "Mining Location-Based Social Networks for Criminal Activity Prediction".24th Wireless and Optical Communication Conference (WOCC), 2015, pp. 185-189.
[26]V. Ferreira, A. Valejo, P. Valdivia and J. Valverde-Rebaza, "Exploiting Geographical Data to improve Recommender Systems for Business Opportunities in Urban Areas."8th Brazilian Conference on Intelligent Systems (BRACIS), 2019, pp. 568-573.
[27]W. G. R. M. P. S. Rathnayake, "Google Maps Based Travel Planning & Analyzing System (TPAS)."International Conference on Current Trends towards Converging Technologies (ICCTCT), 2018, pp. 1-5.
[28]T. Otsuka, R. Suzuki, S. Kawaguchi and T. Ito, "An Implementation of a Geolocation Information-Sharing System. "IEEE International Conference on Service-Oriented Computing and Applications (SOCA) ,2011, pp. 1-2.
[29]K. Yamamoto and P. Vařacha, "Sightseeing Navigation System for Foreign Tourists in Japanese Urban Area. "Smart City Symposium Prague (SCSP), 2019, pp. 1-6.
[30]K. Yamamoto and T. Ikeda,, T, "Social Recommendation GIS for Urban Tourist Spots."18th International Conference on High Performance Computing and Communications; IEEE 14th International Conference on Smart City; IEEE 2nd International Conference on Data Science and Systems (HPCC/SmartCity/DSS), 2016, pp. 50-57.
[31]Z. Gou et al, A, "An interest-based tour planning tool by process mining from Twitter. "IEEE 5th Global Conference on Consumer Electronics, 2016, pp. 1-4.
[32]S. Chang, H. Huang, Y. Huang, C. Yang, C. Hsu and J. Chen, "An Efficient Geographical Place Mining Strategy for Social Networking Services." IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW), 2019, pp. 1-2.
[33]J. Huang, Y. Wei, J. Yi and M. Liu, "An Improved kNN Based on Class Contribution and Feature Weighting."10th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA), 2018, pp. 313-316.
[34]S. Xu, Q. Luo, H. Li and L. Zhang, "Time Series Classification Based on Attributes Weighted Sample Reducing KNN."Second International Symposium on Electronic Commerce and Security, 2009, Vol. 2, pp. 194-199.
[35]F. Sanei, A. Harifi and S. Golzari, "Improving the precision of KNN classifier using nonlinear weighting method based on the spline interpolation."7th International Conference on Computer and Knowledge Engineering (ICCKE), 2017, pp. 289-292.
[36]S. Liu, P. Zhu, and S. Qin, "An Improved Weighted KNN Algorithm for Imbalanced Data Classification."IEEE 4th International Conference on Computer and Communications (ICCC), 2018, pp. 1814-1819.
[37]S. Liu and Q. Xiao, "A Signer-Independent Sign Language Recognition System Based on the Weighted KNN/HMM," 7th International Conference on Intelligent Human-Machine Systems and Cybernetics, Hangzhou, 2015, pp. 186-189.
[38]L. Cheng, Y. Li, M. Zhang and C. Wang, "A Fingerprint Localization Method Based on Weighted KNN Algorithm," IEEE 18th International Conference on Communication Technology (ICCT), Chongqing, 2018, pp. 1271-1275.
[39]Z. Chen, B. Li and B. Han, "Improve Regression Accuracy by using an Attribute Weighted KNN Approach," 13th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD), Guilin, 2017, pp. 1838-1843.
[40]Z. Liu, X. Luo and T. He, "Indoor Positioning System based on the Improved W-KNN Algorithm," IEEE 2nd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), Chongqing, 2017, pp. 1355-1359.
[41]Antonio, Franklin, Faster line segment intersection. Graphics Gems III (IBM Version): Morgan Kaufmann, 1992, pp. 199-202.
[42]C. J. Wang, C. Y. Ng, R. H. Brook, "Response to COVID-19 in Taiwan: Big Data Analytics, New Technology, and Proactive Testing." JAMA, Vol. 323, No. 14, 2020, pp. 1341–1342.
[43]F. Rosenblatt, "The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain, "Psychological review, 1958, Vol. 65, No. 6, pp. 386-408.
[44]C. Dong, D. S. Yeung, "2-Stage Instance Selection Algorithm for KNN based on Nearest Unlike Neighbors," International Conference on Machine Learning and Cybernetics, Qingdao, China, 2010, pp. 11-14.
[45]E. Kolman and M. Margaliot, "Knowledge Extraction from Neural Networks using the All-Permutations Fuzzy Rule Base: the LED Display Recognition Problem," IEEE Transactions on Neural Networks, 2007, Vol. 18, No. 3, pp. 925-931.
[46]R. Xin, J. Zhang and Y. Shao, "Complex Network Classification with Convolutional Neural Network," Tsinghua Science and Technology, 2020, Vol. 25, No. 4, pp. 447-457.
[47]Colaboratory, https://research.google.com/colaboratory/faq.html ，查詢日期:2020年6月.
[48]K. Jun, D. Lee, K. Lee, S. Lee and M. S. Kim, "Feature Extraction Using an RNN Autoencoder for Skeleton-Based Abnormal Gait Recognition," IEEE Access, 2020, Vol. 8, pp. 19196-19207.
[49]S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory," Neural Computation, 1997, Vol. 9, No. 8, pp. 1735-1780.
[50]I. Kusumaningtyas and G. O. F. Parikesit, "Digital Image Analysis of Wayang Kulit Movements," 8th International Conference on Information Technology and Electrical Engineering (ICITEE), Yogyakarta, 2016, pp. 1-4.
[51]GPS,https://zh.wikipedia.org/wiki/%E5%85%A8%E7%90%83%E5%AE%9A%E4%BD%8D%E7%B3%BB%E7%BB%9F , 查詢日期:2020年6月.
[52]Z. Cao, T. Simon, S. Wei and Y. Sheikh, "Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, 2017, pp. 7291-7299.
[53]A. Viswakumar, V. Rajagopalan, T. Ray and C. Parimi, "Human Gait Analysis Using OpenPose," Fifth International Conference on Image Information Processing (ICIIP), Shimla, India, 2019, pp. 310-314.
[54]5G, https://zh.wikipedia.org/wiki/5G 查詢日期:2020年6月.
[55]D. P. Kingma and B. Jimmy, "Adam: A Method for Stochastic Optimization,"The 3rd International Conference for Learning Representations, San Diego, 2015.