人工物料搬運是造成肌肉骨骼傷病中下背痛的其中一個主要原因，不當的抬舉動作易造成腰椎第五節與骶骨第一節(L5/S1)的關節處受傷，因此評估人工抬舉作業中L5/S1關節負荷情形是很重要的。在過去評估下背負荷的研究多選擇在實驗室環境中以動作追蹤系統與測力板進行量測，並搭配生物力學模型中的Bottom-up模型來計算L5/S1的受力與力矩。然而，動作追蹤系統與測力板雖較精準，但基於工作場地、成本等限制較不適合實際應用於作業環境中。因此，許多學者試圖發展利用便宜且便攜的深度相機的自動骨架圖進行肌肉骨骼傷病的相關評估。雖然許多不同領域的研究指出深度相機在人因評估上有一定的準確度，然而其中以深度相機自動骨架圖的數據做為下背關節位置時的誤差較高，更遑論利用此數據來進行準確的下背負荷評估。
過去鮮少研究推算以深度相機數據為基礎的真實下背關節位置，且透過其來計算人工抬舉作業下L5/S1的負荷情形並評估其準確度。人工神經網路可透過學習所給予的輸入與目標參數以建立預測模型。隨著相關技術的日趨成熟，透過人工神經網路來預測以深度相機為基礎的下背關節位置，有望改善深度相機偵測下背關節誤差的問題。
綜合上述，本研究設定光學式動作追蹤系統數據，透過專業生物力學分析軟體所定義的L5/S1關節位置為目標參數、以深度相機偵測抬舉過程關節角度與受測者生理資訊等數據做為輸入參數，開發一基於人工神經網路之深度相機評估人工抬舉時下背負荷的預測模型，並評估其表現。所建立的神經網路可透過深度相機數據來預測L5/S1關節的三維座標位置，搭配同樣由深度相機所偵測的上肢關節數據，即可在不需要動作追蹤系統與測力板的情形下，以Top-down模型計算出人員在對稱抬舉動作之下的下背負荷數據。
本研究以套用光學式動作追蹤系統與測力板數據的Bottom-up模型所得的L5/S1力矩與壓力做為標準，比較利用所開發模型的預測結果或以深度相機自動骨架數據在計算不同人工對稱抬舉動作—地板到肩高(FS)、地板到指節高(FK)、指節高到肩高(KS)之下背力矩、壓力的誤差與數據一致性。
實驗結果顯示，以平均絕對誤差(Mean absolute error, MAE)來看，基於人工神經網路之深度相機（最大總力矩MAE: 21.04 Nm）預測的L5/S1關節位置整體來說較深度相機自動骨架圖（最大總力矩MAE: 55.82 Nm）位置準確，且應用於下背負荷之評估也有較好的表現。綜合組內相關係數(Intra-class correlation coefficient, ICC)分析整體動作之平均與最大下背力矩的結果顯示，基於人工神經網路之深度相機在預測抬舉過程的矢狀面平均力矩(MAE: 12.20 Nm, ICC: 0.852)、平均總力矩(MAE: 13.68 Nm, ICC: 0.808)、最大總力矩(MAE: 21.04 Nm, ICC: 0.784)與平均壓力(MAE: 192.44 N, ICC: 0.835)均有極佳的一致性結果(ICC > 0.75)，並且整體來說優於以深度相機自動骨架圖所計算的結果。
雖然基於人工神經網路之深度相機預測的下背負荷仍與標準數據有差異，但在較不適合使用動作追蹤系統與測力板的追蹤環境下，本研究所開發之方法仍能協助提高單純使用深度相機原始數據作為抬舉作業下背負荷評估的準確度。

Manual material handling (MMH) is one of main causes of low back pain (LBP) in musculoskeletal injuries. Incorrect lifting postures can easily cause injury of the joint between the fifth lumbar vertebrae and first sacral vertebra (L5/S1). Therefore, it is very important to assess L5/S1 load in MMH tasks. In the past, most studies measured low back load in a laboratory environment with motion tracking systems and force plates and used bottom-up model to evaluate the force and moment in L5/S1 joint. Although motion tracking systems and force plates have high accuracy, due to space and cost restrictions, such laboratory equipment may not be suitable for application in a real workplace. Therefore, many researchers have applied skeleton data identified by portable and low-cost depth cameras for musculoskeletal evaluation and several studies have indicated that a depth camera is sufficiently accurate for ergonomic assessment. Nevertheless, errors in low back joint position identified by the depth camera are relatively high, apart from assessing the low back load. In the past, few if any studies estimated the true low back joint position based on depth camera skeleton data and used it to evaluate L5/S1 load and accuracy during manual lifting tasks. An artificial neural network (ANN) can build a prediction model from network input and target parameters. Accordingly, ANN is a promising tool to predict low back joint position based on a depth camera and reduce errors in joint assessment. This study took L5/S1 position data extracted from motion tracking systems and defined in biomechanical analysis software as target parameters, and the joint angle identified by the depth camera during a lifting task and subject physiological information, etc., as input parameters to develop an artificial neural network-based depth camera predictive model. The established model successfully predicted 3D coordinates of the L5/S1 joint position through depth camera data. For the upper limb joint data identified by the depth camera, top-down model was used to calculate low back load during lifting without using motion tracking systems and force plates. As reference, L5/S1 moment and compressive force were calculated from optical motion tracking systems and force plates data by using bottom-up model. The validity and error of the developed model and depth camera skeleton data were evaluated by comparing L5/S1 moment and force in three different symmetric lifting motions: floor to shoulder height (FS), floor to knuckle height (FK), and knuckle to shoulder height (KS) and compared with the reference data, respectively. Results for mean absolute error (MAE) showed that the artificial neural network-based depth camera (peak total moment MAE: 21.04 Nm) had better accuracy in predicting L5/S1 joint position and low back load than depth camera skeleton data (peak total moment MAE: 55.82 Nm). Intra-class correlation coefficient (ICC) revealed that the mean sagittal plane moment (MAE: 12.20 Nm, ICC: 0.852), mean total moment (MAE: 13.68 Nm, ICC: 0.808), peak total moment (MAE: 21.04 Nm, ICC: 0.784) and mean compressive force (MAE: 192.44 N, ICC: 0.835) in low back had excellent reliability (ICC > 0.75) in the proposed method. Overall, the artificial neural network-based depth camera had better results than the depth camera skeleton data. The low back load estimated by the artificial neural network-based depth camera still had some errors compared to the reference data. In sum, in circumstances where use of motion tracking systems and force plates is not possible, the proposed method improves the accuracy of low back load assessment in lifting when using a depth camera.

致謝 I
摘要 II
Abstract IV
目錄 VI
圖目錄 IX
表目錄 XI
第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 4
1.3 研究架構 5
第二章 文獻探討 7
2.1 肌肉骨骼傷病 7
2.1.1 肌肉骨骼傷病調查 7
2.1.2 肌肉骨骼傷病與針對下背痛之研究 7
2.2 動作分析儀器與軟體 11
2.2.1 動作追蹤系統與其應用 11
2.2.2 測力板與其應用 12
2.2.3 動作分析軟體 14
2.3 逆動力學模型與應用 15
2.3.1 逆動力學模型 15
2.3.2 Bottom-up模型與Top-down模型 17
2.3.3 逆動力學模型之應用 18
2.4 下背壓力 19
2.5 深度相機 23
2.5.1 深度相機的應用 23
2.5.2 影響深度相機偵測的因子 25
2.6 人工神經網路與應用 28
2.6.1 人工神經網路 28
2.6.2 人工神經網路之應用 30
2.7 小結 33
第三章 研究方法 35
3.1 實驗儀器與設備 35
3.2 實驗設計與流程 38
3.2.1 自變項與依變項 38
3.2.2 研究參與者 39
3.2.3 環境架設 39
3.2.4 實驗流程 40
3.3 數據分析 43
3.3.1 數據預處理 45
3.3.2 人體骨骼模型建立與測力板的匯入 45
3.3.3 人工神經網路之建立 47
3.3.4 模型驗證 49
3.3.5 下背力矩之計算 50
3.3.6 下背壓力計算 54
3.4 統計分析 55
第四章 結果 56
4.1 人工神經網路結果 57
4.2 下背力矩結果 66
4.2.1 敘述性統計 66
4.2.2 組內相關係數 74
4.3 下背壓力結果 78
4.3.1 敘述性統計 78
4.3.2 組內相關係數 82
第五章 討論 84
5.1 人工神經網路 84
5.2 ANN-KinectTM與KinectTM之探討 86
5.2.1 下背力矩 86
5.2.2 下背壓力 91
5.3 ANN-KinectTM相機與參考標準之探討 94
5.3.1 下背力矩 94
5.3.2 下背壓力 98
5.4 ANN-KinectTM與過去評估方法之比較 100
5.5 研究限制 101
第六章 結論 102
參考文獻 103

1. Abobakr, A., Nahavandi, D., Hossny, M., Iskander, J., Attia, M., Nahavandi, S. & Smets, M. (2019). RGB-D ergonomic assessment system of adopted working postures. Applied ergonomics, 80: 75-88.
2. Aghazadeh, F., Arjmand, N. & Nasrabadi, A. (2019). Coupled artificial neural networks to estimate 3D whole-body posture, lumbosacral moments, and spinal loads during load-handling activities. Journal of Biomechanics: 109332.
3. Ahmed, M. H. (2016). Kinect-based human gait recognition using static and dynamic features. International journal of computer science information security
14(12): 425.
4. Alkronz, E. S., Moghayer, K. A., Meimeh, M., Gazzaz, M., Abu-Nasser, B. S. & Abu-Naser, S. S. (2019). Prediction of Whether Mushroom is Edible or Poisonous Using Back-propagation Neural Network.
5. Allegri, M., Montella, S., Salici, F., Valente, A., Marchesini, M., Compagnone, C., Baciarello, M., Manferdini, M. E. & Fanelli, G. (2016). Mechanisms of low back pain: a guide for diagnosis and therapy. F1000Research, 5.
6. Amita, J., Singh, J. S. & Kumar, G. P. (2015). Prediction of bus travel time using artificial neural network. International Journal for Traffic and Transport Engineering, 5(4): 410-424.
7. Arjmand, N., Amini, M., Shirazi-Adl, A., Plamondon, A. & Parnianpour, M. (2015). Revised NIOSH Lifting Equation May generate spine loads exceeding recommended limits. International journal of industrial ergonomics, 47: 1-8.
8. Awwalu, J. & Nonyelum, O. (2019). On Holdout and Cross Validation A Comparison between Neural Network and Support Vector Machine. International Journal of Trend in Research and Development, 6(2): 235-239.
9. Behjati, M. & Arjmand, N. (2018). Biomechanical Assessment of the NIOSH Lifting Equation in Asymmetric Load-Handling Activities Using a Detailed Musculoskeletal Model. Human factors.
10. Bell, A. L., Brand, R. A. & Pedersen, D. R. (1989). Prediction of hip joint centre location from external landmarks. Human movement science, 8(1): 3-16.
11. Bell, A. L., Pedersen, D. R. & Brand, R. A. (1990). A comparison of the accuracy of several hip center location prediction methods. Journal of biomechanics, 23(6): 617-621.
12. Bhattacharya, A. (2014). Costs of occupational musculoskeletal disorders (MSDs) in the United States. International journal of industrial ergonomics, 44(3): 448-454.
13. Bhattacharya, A., McGlothlin, J.D (2012). Occupational ergonomics: theory and applications. CRC Press.
14. Bickley, C., Linton, J., Sullivan, E., Mitchell, K., Slota, G. & Barnes, D. (2019). Comparison of simultaneous static standing balance data on a pressure mat and force plate in typical children and in children with cerebral palsy. Gait & posture, 67: 91-98.
15. Bobbert, M. F. a. G. J. v. I. S. (1988). Coordination in vertical jumping. Journal of biomechanics, 21: 249-262.
16. Bonnechere, B., Jansen, B., Salvia, P., Bouzahouene, H., Sholukha, V., Cornelis, J., Rooze, M. & Van Sint Jan, S. (2014). Determination of the precision and accuracy of morphological measurements using the Kinect™ sensor: Comparison with standard stereophotogrammetry. Ergonomics, 57(4): 622-631.
17. Brodie, M. A., Walmsley, A. & Page, W. (2008). The static accuracy and calibration of inertial measurement units for 3D orientation. Computer methods in biomechanics and biomedical engineering, 11(6): 641-648.
18. Buchanan, T. S., Lloyd, D. G., Manal, K. & Besier, T. F. (2005). Estimation of muscle forces and joint moments using a forward-inverse dynamics model. Medicine & science in sports & exercise, 37(11): 1911-1916.
19. Bullas, A. M., Choppin, S., Heller, B. & Wheat, J. (2016). Validity and repeatability of a depth camera-based surface imaging system for thigh volume measurement. Journal of sports sciences, 34(20): 1998-2004.
20. CCOHS. (2020). Musculoskeletal disorders (MSD) prevention manual. Retrieved January 14, 2020, from https://www.ccohs.ca/products/publications/msd/.
21. Chaffin, D. B. (1969). A computerized biomechanical model—development of and use in studying gross body actions. Journal of biomechanics, 2(4): 429-441.
22. Chang, C. C., Hsiang, S., Dempsey, P. G. & McGorry, R. W. (2003). A computerized video coding system for biomechanical analysis of lifting tasks. International Journal of Industrial Ergonomics, 32(4): 239-250.
23. Chang, Y.-J., Chen, S.-F. & Huang, J.-D. (2011). A Kinect-based system for physical rehabilitation: A pilot study for young adults with motor disabilities. Research in developmental disabilities, 32(6): 2566-2570.
24. Clark, R. A., Bower, K. J., Mentiplay, B. F., Paterson, K. & Pua, Y. H. (2013). Concurrent validity of the Microsoft Kinect for assessment of spatiotemporal gait variables. Journal of biomechanics, 46(15): 2722-2725.
25. Cleather, D. J. & Bull, A. M. (2010). Influence of inverse dynamics methods on the calculation of inter-segmental moments in vertical jumping and weightlifting. Biomedical Engineering Online, 9: 74.
26. Coenen, P., Kingma, I., Boot, C. R., Twisk, J. W., Bongers, P. M. & van Dieen, J. H. (2013). Cumulative low back load at work as a risk factor of low back pain: a prospective cohort study. Journal of occupatioanl rehabilitation, 23(1): 11-18.
27. Coyte, P. C., Asche, C. V., Croxford, R. & Chan, B. (1998). The economic cost of musculoskeletal disorders in Canada. Official Journal of the American College of Rheumatology, 11(5): 315-325.
28. Cremasco, M., Giustetto, A., Caffaro, F., Colantoni, A., Cavallo, E. & Grigolato, S. (2019). Risk Assessment for Musculoskeletal Disorders in Forestry: A Comparison between RULA and REBA in the Manual Feeding of a Wood-Chipper. International journal of environment research and public health, 16(5).
29. Davis III, R. B., Ounpuu, S., Tyburski, D. & Gage, J. R. (1991). A gait analysis data collection and reduction technique. Human movement science, 10(5): 575-587.
30. Davis, K. & Marras, W. (2005). Load spatial pathway and spine loading: how does lift origin and destination influence low back response? Ergonomics, 48(8): 1031-1046.
31. De Leva, P. (1996). Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters. Journal of biomechanics, 29(9): 1223-1230.
32. Delp, S. L., Anderson, F. C., Arnold, A. S., Loan, P., Habib, A., John, C. T., Guendelman, E. & Thelen, D. G. (2007). OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Transaction on biomedical engineering, 54(11): 1940-1950.
33. Dempsey, P. G., Hashemi & Lobat (1999). Analysis of workers' compensation claims associated with manual materials handling. Ergonomics, 42(1): 183-195.
34. Eltoukhy, M., Kuenze, C., Andersen, M. S., Oh, J. & Signorile, J. (2017). Prediction of ground reaction forces for Parkinson's disease patients using a kinect-driven musculoskeletal gait analysis model. Medical engineering and physics, 50: 75-82.
35. Faber, G. S., Chang, C. C., Kingma, I. & Dennerlein, J. T. (2013). Estimating dynamic external hand forces during manual materials handling based on ground reaction forces and body segment accelerations. Journal of biomechanics, 46(15): 2736-2740.
36. Faber, G. S., Chang, C. C., Kingma, I., Dennerlein, J. T. & van Dieen, J. H. (2016). Estimating 3D L5/S1 moments and ground reaction forces during trunk bending using a full-body ambulatory inertial motion capture system. Journal of biomechanics, 49(6): 904-912.
37. Faber, G. S., Koopman, A., Kingma, I., Chang, C., Dennerlein, J. & van Dieën, J. (2018). Continuous ambulatory hand force monitoring during manual materials handling using instrumented force shoes and an inertial motion capture suit. Journal of biomechanics, 70: 235-241.
38. Fathallah, F. A., Marras, W. S. & Parnianpour, M. (1999). Regression models for predicting peak and continuous three-dimensional spinal loads during symmetric and asymmetric lifting tasks. Human factors, 41(3): 373-388.
39. Fleiss, J. L. (2011). Design and analysis of clinical experiments, John Wiley & Sons.
40. Fushiki, T. (2011). Estimation of prediction error by using K-fold cross-validation. Statistics and Computing, 21(2): 137-146.
41. Gagnon, D., Arjmand, N., Plamondon, A., Shirazi-Adl, A. & Larivière, C. (2011). An improved multi-joint EMG-assisted optimization approach to estimate joint and muscle forces in a musculoskeletal model of the lumbar spine. Journal of biomechanics, 44(8): 1521-1529.
42. Galna, B., Barry, G., Jackson, D., Mhiripiri, D., Olivier, P. & Rochester, L. (2014). Accuracy of the Microsoft Kinect sensor for measuring movement in people with Parkinson's disease. Gait & posture, 39(4): 1062-1068.
43. Garg, A. & Moore, J. (1992). Epidemiology of low-back pain in industry. Occupational medicine, 7(4): 593-608.
44. Gholipour, A. & Arjmand, N. (2016). Artificial neural networks to predict 3D spinal posture in reaching and lifting activities; Applications in biomechanical models. Journal of biomechanics, 49(13): 2946-2952.
45. Grant, K. A., Johnson, P. W. & Galinsky, T. L. (1995). Evaluation of an accelerometric activity monitor as an exposure assessment tool in ergonomic studies. Applied Occupational and Environmental Hygiene, 10(5): 461-466.
46. Harsted, S., Holsgaard-Larsen, A., Hestbaek, L., Boyle, E. & Lauridsen, H. H. (2019). Concurrent validity of lower extremity kinematics and jump characteristics captured in pre-school children by a markerless 3D motion capture system. Chiropractic & manual therapies, 27: 39.
47. Hendershot, B. D. & Wolf, E. J. (2014). Three-dimensional joint reaction forces and moments at the low back during over-ground walking in persons with unilateral lower-extremity amputation. Clinical biomechanics, 29(3): 235-242.
48. Hof, A. L. (1992). An explicit expression for the moment in multibody systems. Journal of Biomechanics, 25(10): 1209-1211.
49. Hoogendoorn, W.E., Bongers & M., P. (2000). Flexion and rotation of the trunk and lifting at work are risk factors for low back pain: results of a prospective cohort study. Spine, 25(23): 3087-3092.
50. Hoozemans, M. J., Kingma, I., de Vries, W. H. & van Dieën, J. H. (2008). Effect of lifting height and load mass on low back loading. Ergonomics, 51(7): 1053-1063.
51. Hoozemans, M. J., Kuijer, P. P. F., Kingma, I., van Dieën, J. H., de Vries, W. H., van der Woude, L. H., Veeger, D. J., van der Beek, A. J. & Frings-Dresen, M. H. (2004). Mechanical loading of the low back and shoulders during pushing and pulling activities. Ergonomics, 47(1): 1-18.
52. Hornik, K., Stinchcombe, M. & White, H. (1990). Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks. Neural networks, 3(5): 551-560.
53. Hoy, D., March, L., Brooks, P., Blyth, F., Woolf, A., Bain, C., Williams, G., Smith, E., Vos, T., Barendregt, J., Murray, C., Burstein, R. & Buchbinder, R. (2014). The global burden of low back pain: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis, 73(6): 968-974.
54. Hush, D. R. & Horne, B. G. (1993). Progress in supervised neural networks. IEEE signal processing magazine, 10(1): 8-39.
55. Kadaba, M. P., Ramakrishnan, H. & Wootten, M. (1990). Measurement of lower extremity kinematics during level walking. Journal of orthopaedic research
8(3): 383-392.
56. Kang, S. Y. (1992). An investigation of the use of feedforward neural networks for forecasting.
57. Karhu, O., Kansi, P. & Kuorinka, I. (1977). Correcting working postures in industry: a practical method for analysis. Applied ergonomics, 8(4): 199-201.
58. Keyserling, W., Stetson, D., Silverstein, B. & Brouwer, M. (1993). A checklist for evaluating ergonomic risk factors associated with upper extremity cumulative trauma disorders. Ergonomics, 36(7): 807-831.
59. Keyserling, W. M., Brouwer, M. & Silverstein, B. A. (1992). A checklist for evaluating ergonomic risk factors resulting from awkward postures of the legs, trunk and neck. International Journal of Industrial Ergonomics, 9(4): 283-301.
60. Kingma, I., de Looze, M. P., Toussaint, H. M., Klijnsma, H. G. & Bruijnen, T. B. (1996). Validation of a full body 3-D dynamic linked segment model. Human movement science, 15(6): 833-860.
61. Kipp, K., Giordanelli, M. & Geiser, C. (2018). Predicting net joint moments during a weightlifting exercise with a neural network model. Journal of Biomechanics, 74: 225-229.
62. Kirking, B., El-Gohary, M. & Kwon, Y. (2016). The feasibility of shoulder motion tracking during activities of daily living using inertial measurement units. Gait posture, 49: 47-53.
63. Kitsikidis, A., Dimitropoulos, K., Douka, S. & Grammalidis, N. (2014). Dance analysis using multiple kinect sensors. 2014 International conference on computer vision theory and applications, 2: 789-795.
64. Koopman, A. S., Kingma, I., Faber, G. S., Bornmann, J. & van Dieen, J. H. (2018). Estimating the L5S1 flexion/extension moment in symmetrical lifting using a simplified ambulatory measurement system. Journal of biomechanics, 70: 242-248.
65. Kordanuli, B., Barjaktarović, L., Jeremić, L. & Alizamir, M. (2017). Appraisal of artificial neural network for forecasting of economic parameters. Physica A: Statistical Mechanics and its Applications, 465: 515-519.
66. Kotsifaki, A., Whiteley, R. & Hansen, C. (2018). Dual Kinect v2 system can capture lower limb kinematics reasonably well in a clinical setting: concurrent validity of a dual camera markerless motion capture system in professional football players. BMJ open sport & exercise medicine, 4(1).
67. Kuorinka, I., Jonsson, B., Kilbom, A., Vinterberg, H., Biering-Sørensen, F., Andersson, G. & Jørgensen, K. (1987). Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Applied egonomics, 18(3): 233-237.
68. Lafta, H. A., Hasan, Z. F. & Ayoob, N. K. (2019). Classification of medical datasets using back propagation neural network powered by genetic-based features elector. International Journal of Electrical and Computer Engineering, 9(2): 1379.
69. Larivière, C. & Gagnon, D. (1998). Comparison between two dynamic methods to estimate triaxial net reaction moments at the L5/S1 joint during lifting. Clinical Biomechanics, 13(1): 36-47.
70. Lippmann, R. (1987). An introduction to computing with neural nets. IEEE Assp magazine, 4(2): 4-22.
71. Loconsole, C., Barbosa, N., Frisoli, A. & Orvalho, V. C. (2012). A new marker-less 3d kinect-based system for facial anthropometric measurements. International Conference on Articulated Motion and Deformable Objects, Springer.
72. Manghisi, V. M., Uva, A. E., Fiorentino, M., Bevilacqua, V., Trotta, G. F. & Monno, G. (2017). Real time RULA assessment using Kinect v2 sensor. Applied ergonomics, 65: 481-491.
73. Marras, W. S., Lavender, S. A., Ferguson, S. A., Splittstoesser, R. E., Yang, G. & Schabo, P. (2010). Instrumentation for measuring dynamic spinal load moment exposures in the workplace. Journal of electromyography and kinesiology, 20(1): 1-9.
74. Marras, W. S., Lavender, S. A., Leurgans, S. E., Fathallah, F. A., Ferguson, S. A., Allread, W. G. & Rajulu, S. L. (1995). Biomechanical risk factors for occupationally related low back disorders. Ergonomics, 38(2): 377-410.
75. Matthew, R. P., Seko, S., Bajcsy, R. & Lotz, J. (2018). Kinematic and kinetic validation of an improved depth camera motion assessment system using rigid bodies. IEEE journal of biomedical and health informatics, 23(4): 1784-1793.
76. McAtamney, L. & Corlett, E. N. (1993). RULA: a survey method for the investigation of work-related upper limb disorders. Applied ergonomics, 24(2): 91-99.
77. McClelland, J. L. & Rumelhart, D. E. (1989). Explorations in parallel distributed processing: A handbook of models, programs, and exercises, MIT press.
78. McGill, S. M., NORMAN, R. W. & CHOLEWICKI, J. (1996). A simple polynomial that predicts low-back compression during complex 3-D tasks. Ergonomics, 39(9): 1107-1118.
79. Mehrizi, R., Xu, X., Zhang, S., Pavlovic, V., Metaxas, D. & Li, K. (2017). Using a marker-less method for estimating L5/S1 moments during symmetrical lifting. Applied Ergonomics, 65: 541-550.
80. Muller, B., Ilg, W., Giese, M. A. & Ludolph, N. (2017). Validation of enhanced kinect sensor based motion capturing for gait assessment. PLoS One, 12(4).
81. Naeemabadi, M., Dinesen, B., Andersen, O. K. & Hansen, J. (2018). Investigating the impact of a motion capture system on Microsoft Kinect v2 recordings: A caution for using the technologies together. PloS one, 13(9).
82. Nasr-Esfahani, E., Samavi, S., Karimi, N., Soroushmehr, S. M. R., Jafari, M. H., Ward, K. & Najarian, K. (2016). Melanoma detection by analysis of clinical images using convolutional neural network. 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), IEEE.
83. Nelson, N. A. & Hughes, R. E. (2009). Quantifying relationships between selected work-related risk factors and back pain: a systematic review of objective biomechanical measures and cost-related health outcomes. International journal of industrial ergonomics, 39(1): 202-210.
84. Ngoh, K. J.-H., Gouwanda, D., Gopalai, A. A. & Chong, Y. Z. (2018). Estimation of vertical ground reaction force during running using neural network model and uniaxial accelerometer. Journal of biomechanics, 76: 269-273.
85. Ning, X. & Guo, G. (2012). Assessing spinal loading using the Kinect depth sensor: A feasibility study. IEEE Sensors journal, 13(4): 1139-1140.
86. Norasi, H., Koenig, J. & Mirka, G. A. (2019). The effects of load weight and load starting height on variability of lifting kinematics and kinetics. International journal of industrial ergonomics, 73.
87. Otte, K., Kayser, B., Mansow-Model, S., Verrel, J., Paul, F., Brandt, A. U. & Schmitz-Hubsch, T. (2016). Accuracy and Reliability of the Kinect Version 2 for Clinical Measurement of Motor Function. PLoS One, 11(11).
88. Pagliari, D. & Pinto, L. (2015). Calibration of kinect for xbox one and comparison between the two generations of microsoft sensors. Sensors, 15(11): 27569-27589.
89. Pastor, I., Hayes, H. A. & Bamberg, S. J. (2012). A feasibility study of an upper limb rehabilitation system using kinect and computer games. 2012 annual international conference of the ieee engineering in medicine and biology society: 1286-1289.
90. Pfister, A., West, A. M., Bronner, S., Noah, J. A. & technology (2014). Comparative abilities of Microsoft Kinect and Vicon 3D motion capture for gait analysis. Journal of medical engineering technology, 38(5): 274-280.
91. Picard, R. R. & Berk, K. N. (1990). Data splitting. The American Statistician, 44(2): 140-147.
92. Plamondon, A., Gagnon, M. & Desjardins, P. J. C. B. (1996). Validation of two 3-D segment models to calculate the net reaction forces and moments at the L 5 S 1 joint in lifting. Clinical biomechanics 11(2): 101-110.
93. Plantard, P., Auvinet, E., Pierres, A.-S. L. & Multon, F. (2015). Pose estimation with a kinect for ergonomic studies: Evaluation of the accuracy using a virtual mannequin. Sensors, 15(1): 1785-1803.
94. Plantard, P., Shum, H. P. & Multon, F. (2017). Filtered pose graph for efficient kinect pose reconstruction. Multimedia tools applications, 76(3): 4291-4312.
95. Rab, G., Petuskey, K. & Bagley, A. (2002). A method for determination of upper extremity kinematics. Gait & posture, 15(2): 113-119.
96. Rane, A. & Kumar, A. (2018). Sentiment classification system of Twitter data for US airline service analysis. 2018 IEEE 42nd Annual Computer Software and Applications Conference (COMPSAC), IEEE.
97. Rau, H.-H., Hsu, C.-Y., Lin, Y.-A., Atique, S., Fuad, A., Wei, L.-M. & Hsu, M.-H. (2016). Development of a web-based liver cancer prediction model for type II diabetes patients by using an artificial neural network. Computer methods and programs in biomedicine, 125: 58-65.
98. Revett, K., Gorunescu, F., Gorunescu, M., El-Darzi, E. & Ene, M. (2005). A breast cancer diagnosis system: a combined approach using rough sets and probabilistic neural networks. EUROCON 2005-The International Conference on" Computer as a Tool", IEEE.
99. Rosenblatt, F. (1958). The perceptron: a probabilistic model for information storage and organization in the brain. Psychological review, 65(6): 386.
100. Salau, J., Haas, J., Thaller, G., Leisen, M. & Junge, W. (2016). Developing a multi-Kinect-system for monitoring in dairy cows: object recognition and surface analysis using wavelets. Animal, 10(9): 1513-1524.
101. Santos, M. S., Soares, J. P., Abreu, P. H., Araujo, H. & Santos, J. (2018). Cross-validation for imbalanced datasets: Avoiding overoptimistic and overfitting approaches [research frontier]. ieee ComputatioNal iNtelligeNCe magaziNe, 13(4): 59-76.
102. Schmitz, A., Ye, M., Shapiro, R., Yang, R. & Noehren, B. (2014). Accuracy and repeatability of joint angles measured using a single camera markerless motion capture system. Journal of biomechanics, 47(2): 587-591.
103. Seay, J., Selbie, W. S. & Hamill, J. (2008). In vivo lumbo-sacral forces and moments during constant speed running at different stride lengths. Journal of sports sciences, 26(14): 1519-1529.
104. Seo, N. J., Fathi, M. F., Hur, P. & Crocher, V. (2016). Modifying Kinect placement to improve upper limb joint angle measurement accuracy. Journal of hand therapy, 29(4): 465-473.
105. Shotton, J., Fitzgibbon, A., Cook, M., Sharp, T., Finocchio, M., Moore, R., Kipman, A. & Blake, A. (2011). Real-time human pose recognition in parts from single depth images. CVPR 2011, Ieee.
106. Straker, L. M., Smith, A. J., Bear, N., O'Sullivan, P. B. & de Klerk, N. H. (2011). Neck/shoulder pain, habitual spinal posture and computer use in adolescents: the importance of gender. Ergonomics, 54(6): 539-546.
107. Sutthiprapa, S., Vanijja, V. & Likitwon, T. (2017). The deadlift form analysis system using Microsoft Kinect. Procedia computer science, 111: 174-182.
108. Tanaka, R., Takimoto, H., Yamasaki, T. & Higashi, A. (2018). Validity of time series kinematical data as measured by a markerless motion capture system on a flatland for gait assessment. Journal of biomechanics, 71: 281-285.
109. Toussaint, H. M., Commissaris, D. A., Van, D., n, J. H., Reijnen, J. S., Praet, S. F. & Beek, P. J. (1995). Controlling the Ground Reaction Force During Lifting. J Mot Behav, 27(3): 225-234.
110. US Department of Labor (2016). Nonfatal occupational injuries and illnesses resulting in days away from work in 2015.
111. van der Kruk, E. & Reijne, M. M. (2018). Accuracy of human motion capture systems for sport applications; state-of-the-art review. European of journal sport science, 18(6): 806-819.
112. Waters, T. R., Putz-Anderson, V., Garg, A. & Fine, L. J. (1993). Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics, 36(7): 749-776.
113. Wolf, S., Simon, J., Patikas, D., Schuster, W., Armbrust, P. & Doderlein, L. (2008). Foot motion in children shoes: a comparison of barefoot walking with shod walking in conventional and flexible shoes. Gait posture, 27(1): 51-59.
114. Wong, F. S. (1991). Time series forecasting using backpropagation neural networks. Neurocomputing, 2(4): 147-159.
115. Xu, H., Hunt, M., Bo Foreman, K., Zhao, J. & Merryweather, A. (2018). Gait alterations on irregular surface in people with Parkinson's disease. Clinical biomechanics, 57: 93-98.
116. Xu, X., Chang, C.-C., Faber, G. S., Kingma, I. & Dennerlein, J. T. (2012). Estimating 3-D L5/S1 moments during manual lifting using a video coding system: Validity and interrater reliability. Human factors, 54(6): 1053-1065.
117. Xu, X., Chang, C. C. & Lu, M. L. (2012). Two linear regression models predicting cumulative dynamic L5/S1 joint moment during a range of lifting tasks based on static postures. Ergonomics, 55(9): 1093-1103.
118. Xu, X. & McGorry, R. W. (2015). The validity of the first and second generation Microsoft Kinect for identifying joint center locations during static postures. Applied ergonomics, 49: 47-54.
119. Xu, X., Robertson, M., Chen, K. B., Lin, J. H. & McGorry, R. W. (2017). Using the Microsoft Kinect to assess 3-D shoulder kinematics during computer use. Applied egonomics, 65: 418-423.
120. Yamako, G., Chosa, E., Totoribe, K., Fukao, Y. & Deng, G. (2017). Quantification of the sit-to-stand movement for monitoring age-related motor deterioration using the Nintendo Wii Balance Board. PLoS one, 12(11).
121. Yang, L., Zhang, L., Dong, H., Alelaiwi, A. & Saddik, A. E. (2015). Evaluating and Improving the Depth Accuracy of Kinect for Windows v2. IEEE sensors journal, 15(8): 4275-4285.
122. Zehr, J. D., Carnegie, D. R., Welsh, T. N. & Beach, T. A. C. (2018). A comparative analysis of lumbar spine mechanics during barbell- and crate-lifting: implications for occupational lifting task assessments. International journal of occupational safety and ergonomics, 26(1): 1-8.
123. Zennaro, S., Munaro, M., Milani, S., Zanuttigh, P., Bernardi, A., Ghidoni, S. & Menegatti, E. (2015). Performance evaluation of the 1st and 2nd generation Kinect for multimedia applications. 2015 IEEE International Conference on Multimedia and Expo (ICME), IEEE.
124. 邱華慧 (2011). 保母人員工作壓力與肌肉骨骼不適情形之研究. 醫護科技期刊, 13(4): 214-226.
125. 張婓章、張麗秋 (2015). 類神經網路導論-原理與應用. 新北市, 滄海圖書資訊股份有限公司.
126. 張朝昇.(2015). 地勤作業人員肌肉骨骼傷害調查-以台中機場為例. 國立中山醫學大學職業安全衛生學系.
127. 陳致廷.(2017). RGB-D相機應用於動態抬舉作業評估之可行性. 國立清華大學工業工程與工程管理學系.
128. 勞動部勞動及職業安全衛生研究所.(2016). 台灣人體計測及重要計測值.
129. 鍾少軒.(2016). RGB-D相機之人體靜態動作追蹤表現評估. 國立清華大學工業工程與工程管理學系.