臺灣博碩士論文加值系統

數位內容產業的產業規模與產值隨著時間而與日俱增，亞太地區為成長速度最快同
時也是全球第三大市場。台灣的數位內容產業仍有很大的成長空間。數位內容是將各類
素材經過處理後，將資料轉化成數位化格式並賦予新的應用型態；以健康和運動領域來
看，透過數位化技術將資料進一步處理，可以用來研究降低人體傷害以及傷害後的復健
過程。
本研究以拉邦動作分析(LMA)理論為基礎，並將研究範圍鎖定在時間分類的研究分
析，將運動學參數的計算方法以單位時間(Frame)計算，目的在於透過時間的變化來觀察
參數的變化，以改進過去在位移、速度、加速度、角度變化量、角速度、角加速度在計
算上僅利用開始與結束進行分析而容易忽略掉動作中的細微資訊的缺點，計算後將資料
展現透過統計的方法擷取動作特性參數，以RDP 平滑化演算法技術檢測多個門檻值，
再利用統計分類樹的方法找出最適合的門檻值，並找到適合時間分類的動作特性參數，
進行各肢段各質地的動作因子決策準則以及值域的解析，將分析結果儲存入本研究所設
計的動作質地資料庫中，並提供一套框架程序，供後人可以繼續建構完整的動作質地資
料庫，或運用動作質地資料庫操作出特定肢段特定質地的動作。

The scale and output value of digital content industries are steadily on the increase. And
digital content industries still has a very big growth space in Taiwan. Digital content includes
a lot of material information that will be processed into digital format and give new types of
applications. By the medical service and movement domain, further process will be done by
the digitizing technology the material. Afterwards, it may be used to study how to reduce the
bodily harm as well as the injury recovery process.
In this study, based on the LMA(Laban Movement Analysis) theory , the scope of the
study focuses on time analysis of the kinematics parameters by the unit interval/unit time
(frame) computation. The purpose of the analysis is to observe the changes of the parameters
by time change, to improve the displacement, the velocity, the acceleration, the angular
variation, the angular velocity, the angular acceleration in the computation only used in the
beginning and end times. After material computation, it then penetrates the statistical methods
to pick up some movement characteristic parameters. We applied RDP smoothing algorithm
to detect the threshold values, used statistical classification tree method to identify the most
appropriate threshold and to find the movement characteristic parameters with suitable time
classification. We then carried the analytical process on the decision-making criteria and
domain value of movement factors to respective body parts/efforts, and the analytical results
were stored into the movement effort database, which could be used to generate the digitized
movement with specific efforts on some specific body parts. We also designed a framework
for future generations to build the complete movement effort database.

致謝 ············································································································································· I
中文論文提要 ··························································································································· II
英文論文提要 ·························································································································· III
目 次 ······································································································································ IV
圖 次 ······································································································································ VI
表 次 ···································································································································· VII
第一章 緒 論 ···················································································································· 1
第一節 研究背景 ········································································································································· 1
第二節 研究動機與目的 ····························································································································· 2
第三節 名詞操作性定義 ····························································································································· 3
第四節 論文架構流程 ································································································································· 4
第二章 文獻探討 ················································································································ 6
第一節 拉邦動作分析 ································································································································· 6
第二節 LMA 量化分析 ······························································································································· 8
第三節 RAMER-DOUGLAS-PEUCKER 演算法 ····························································································· 14
第四節 分類樹 ··········································································································································· 15
第五節 本章總結 ······································································································································· 16
第三章 研究方法 ·············································································································· 17
第一節 研究架構 ······································································································································· 17
第二節 量化參數計算前處理 ··················································································································· 19
第三節 動作特性參數擷取方法 ··············································································································· 22
第四節 資料平滑化 ··································································································································· 24
第五節 資料分析 ······································································································································· 24
第四章 系統設計及結果分析 ·························································································· 27
第一節 系統架構 ······································································································································· 27
第二節 動作因子量化計算 ······················································································································· 28
第三節 動作特性參數擷取 ······················································································································· 29
第四節 平滑化與分類樹處理 ··················································································································· 31
第五節 分析結果與檢驗 ··························································································································· 35
第六節 動作質地資料庫與應用 ··············································································································· 37
第五章 結論與建議 ·········································································································· 40
第一節 結論 ··············································································································································· 40
第二節 未來方向與建議 ··························································································································· 41
參考文獻 ·································································································································· 42
一、 中文部分 ············································································································································ 42
二、 英文部分 ············································································································································ 43
附錄一 人體定位相關表 ········································································································ 44
附錄二 人體關節定義與光點計算法 ···················································································· 47
附錄三 分析結果 ···················································································································· 48
簡 歷 ······································································································································ 60
著作權聲明 ······························································································································ 61

一、 中文部分
1. 方鄒昭聰 & 楊迪強（2008），運用K-Means 演算法探討人體動作因子之組成-以
LMA 為基礎，2008 管理創新與新願景研討會論文集，臺北。
2. 吳啟聰譯（2002），商用統計學入門與應用 (第 1 版)，臺北：麥格羅希爾。
3. 李宗翰（2008），人體動作擷取是別與資訊傳遞以拉邦動作分析理論為基礎，國立
臺北大學資訊管理研究所碩士論文。
4. 國立臺北藝術大學舞蹈學院動作創意實驗室（2006），經濟部學界開發產業技術計
畫執行報告-人體動作質地分析與肢體情緒數位傳達應用開發三年計畫第一年度執
行成果報告書，計畫編號：95-EC-17-A-02-S1-052，臺北。
5. 國立臺北藝術大學舞蹈學院動作創意實驗室（2007），經濟部學界開發產業技術計
畫執行報告-人體動作質地分析與肢體情緒數位傳達應用開發三年計畫第-4 年度
執行成果報告書，計畫編號：95-EC-17-A-02-S1-052，臺北。
6. 陳五洲 & 黃彥慈（2007），拉邦動作分析論，大專體育 (88)。
7. 楊迪強（2008），運用K-Means 演算法探討人體動作因子之組成─以LMA 為基礎，
國立臺北大學資訊管理研究所碩士論文。
8. 經濟部工業局（2007），2007 數位內容產業年鑑 (第 1 版)，臺北：經濟部。
9. 蔡家昌（2002），應用決策樹歸納法探討台灣行動電話市場區隔，國立臺北大學統
計研究所碩士論文。
10. 羅孟剛（2008），資料採礦應用於人體動作質地分析-以LMA 為基礎，國立臺北大
學資訊管理研究所碩士論文。
二、英文部分
1. Badler, I. N., Chi, D. M., & Chopra, S. (1999). Virtual human animation based on
movement observation and cognitive behavior models. Proceedings of the Computer
Animation Conference, (pp. 128-137).
2. Badler, N. I., & Smoliar, S. W. (1979). Digital Representations of Human Movement.
ACM Computing Surveys , 1 (11), pp. 19-38.
3. Callennec, B. L., & Boulic, R. Robust Kinematic Constraint Detection for Motion Data.
Proceedings of the 2006 ACM SIGGRAPH/Eurograp Symposium on Computer
Animation, (pp. 281-290).
4. Chi, D. M., Costa, M., Zhao, L., & Badler, N. The EMOTE Model for Effort and Shape.
Proceedings of the 27th annual conference on computer graphics and interactive
techniques, (pp. 173-182).
5. David Douglas & Thomas Peucker, (1973) "Algorithms for the reduction of the number
of points required to represent a digitized line or its caricature", The Canadian
Cartographer 10(2), 112-122
6. Fangtsou, C. T., & Li, T. H. (2008). Body Movement Acquisition and Construct Data
Exchange Protocol Based on Body Movement Quality Analysis. Proceedings of the
e-CASE 2008 International Joint Conference on Advances in e-Commerce.
7. Fangtsou, C. T., & Yang, T. C. (2008). About Body Movement Factors by Using
K-Means Algorithm. Proceedings of the e-CASE 2008 International Joint Conference on
Advances in e-Commerce.
8. Ieronutti, L., & Chittaro, L. (2005). A Virtual Human Architecture that Integrates
Kinematic, Physical and Behavioral Aspects to Control H-Anim Characters. Proceedings
of the tenth international conference on 3D Web technology, (pp. 39-48).
9. Nagamatsu, T., Kamahara, J., Iko, T., & Tanaka, N. (2008). One-point calibration gaze
tracking based on eyeball kinematics using stereo cameras. Proceedings of the 2008
symposium on Eye tracking research & applications, (pp. 95-98).
10. Urs Ramer, (1972) "An iterative procedure for the polygonal approximation of plane
curves", Computer Graphics and Image Processing, 1(2), 244-256