台灣的洗腎率在全球中居高不下,又台灣健保全額給付洗腎費用,若能提供腎友們一個更完善的醫療服務,便能夠節省台灣的醫療資源。腎友在正常洗腎前,需確保廔管的血液能夠順暢不栓塞,若是在洗腎前才發現廔管栓塞,不僅讓腎友處於危險的狀況,也要臨時去做血管疏通手術。在現今檢測廔管是否栓塞的技術裡,大多是侵入性及成本很高的檢測,如血管攝影術或是核磁共振,若能早一步讓腎友可以自行檢測,能夠以非侵入式、 簡易的儀器測量其血管有無栓塞,不但能提高腎友的存活率,也能夠節省台灣的醫療資源。
血液在人類的血管流動時,會依照血管厚度、性質不同而有複雜的流體力學,因此當血管阻塞時,血流速度與血管流動的聲音 (以下稱血管音) 會不同。本研究透過測量血液在廔管與血管流動的聲音,找出腎友的廔管是否有栓塞的跡象,並幫助腎友能夠自行檢測廔管是否栓塞。論文中會對於正常人與嚴重阻塞需做疏通手術的腎友進行血管音分析,最後得出三個結論 (1) 電腦種類顯著影響血管音訊號,因此要對於頻域訊號做比例化的轉換,(2) 找出依照阻塞位置的不同,嚴重阻塞腎友有2類嚴重阻塞的現象,(3) 提出兩個指標,可以判斷心跳週期時域訊號是否具有穩定週期性性質,在實際阻塞位置手術後兩個指標皆降低。

Taiwan is known to have the greatest number of dialysis patients total compared to other countries. It is a urgent issue that we have to improve medical services and reduce medical cost. One recent successful research regarding dialysis comes from the Industrial Technology Research Institute (ITRI) in Taiwan, which has created a portable measuring device for vascular sounds (ITRI-PMDV). The device tries to measure whether a patient's vascular stenosis is within the safety levels required before any dialysis treatment. This research focuses on detect and classify vascular sounds collected via the above-mentioned ITRI-PMDV
The proposed approach is based on a series of scientific procedures, including (1) Collect the dialysis patients of severe stenosis on AV-fistula. (2) Pre-processing the signal of vascular sounds. (3) applying independent component analysis (ICA) and so-called FastICA algorithm. (4) analyze the signals collected at arteriovenous anastomosis and venous. In this thesis, we get 3 results. (1) Computer is a main factor affect the vascular sound signal. (2) There is two type of severe stenosis of fistula. (3) Find indice called Cor-index and ICA-index which can detect stenosis on fistula.

致謝.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
摘要. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
英文摘要. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
目錄. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
表目錄. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
圖目錄. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
第1 章緒論1
1.1 研究背景. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 研究動機與目的. . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 研究流程與架構. . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 命名定義與數學符號. . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 論文架構. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
第2 章文獻探討9
2.1 訊號處理. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 傅立葉轉換. . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 時域以及頻域的意義. . . . . . . . . . . . . . . . . . . . . . 10
2.2 獨立成份分析(Independent ComponentAnalysis, ICA) . . . . . . . . 11
2.2.1 獨立成份分析的模型及基本假設. . . . . . . . . . . . . . . . . 11
2.2.2 訊號的非高斯性及獨立性. . . . . . . . . . . . . . . . . . . . 12
2.2.3 FastICA演算法. . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.4 ICA的目標函數式. . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.5 FastICA最佳化演算法及得到獨立成分. . . . . . . . . . . . . 18
2.3 分類與預測. . . . . . . . . . . . . . . . . . .. . . . . . . . . 19
2.3.1 決策樹. . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3.2 資訊增益比(Information Gain Ratio) . . . . . . . . . . . . . . 20
2.3.3 決策樹分類模型的績效指標. . . . . . . . . . . . . . . . . . . 22
2.4 血管音資料分析. . . . . . . . . . . . . . . . . .. . . . . . . . 23
第3 章血管音收集位置及資料前處理24
3.1 收集訊號位置及收音機械. . . . . . . . . . . . . . .. . . . . . . 24
3.2 資料前處理. . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.1 設計濾波器. . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2.2 訊號萃取. . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.3 訊號平滑. . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.4 時頻域轉換. . . . . . . . . . . . . . . . . . . . . . . . . . 29
第4 章探討影響血管音量測之顯著因子32
4.1 設計實驗因子表. . . . . . . . . . . . . . . . . . . .. . . . . . 32
4.2 因子實驗分析及結果. . . . . . . . . . . . . . . . . . . . . . . 34
4.3 小結: 電腦種類顯著影響血管音訊號. . . . . . . . . . . . . . . . 34
第5 章探討嚴重阻塞(必須做疏通手術的) 的腎友類型35
5.1 Pair t 檢定分析. . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2 位置1嚴重阻塞的腎友. . . . . . . . . . . . . . . . . . . . . . . 36
5.3 位置3-5嚴重阻塞的腎友. . . . . . . . . . . . . . . . . . . . . . 36
5.4 小結: 有2種嚴重阻塞的腎友. . . . . . . . . . . . . . . . . . . . 38
第6 章本研究提出的反應廔管阻塞的2種指標39
6.1 不同時段的心跳週期訊號. . . . . . . . . . . . . . .. . . . . . . 39
6.2 心跳週期訊號的時域分析. . . . . . . . . . . . . . .. . . . . . . 40
6.2.1 穩定週期性分析流程. . . . . . . . . . . . . . . . . . . .. . . 41
6.2.2 Cor指標. . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6.2.3 ICA指標分析的數學定義. . . . . . . . . . . . . . . . . . . . . 43
6.2.4 ICA指標中的穩定週期性. . . . . . . . . . . . . . . . . . . . . 46
6.2.5 ICA指標中的不穩定週期性. . . . . . . . . . . . . . . . . . . 48
6.3 小結: 2種指標反應嚴重阻塞腎友疏通手術前後的差異. . . . . . . . 50
第7 章探討頻域訊號比時域訊號在分類上何者較佳51
7.1 訊號模擬. . . . . . . . . . . . . . .. . . . . . . . . . . . . . 51
7.2 決定模擬訊號組合. . . . . . . . . . . . . . . . . . . . . . . . 52
7.3 建立分類模型. . . . . . . . . . . . . . . . . . . . . . . . . . 52
7.4 小結: 頻域訊號在分類模型上較佳. . . . . . . . . . . . . . . . . 55
第8 章結論與未來展望58
8.1 結論. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
8.2 未來展望. . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
參考文獻. . . . .. . . . . . . . . . . . . . . . . . . . . . . . .60

[1] USRDS, “2013 USRDS Annual Data Report”, United States Renal Data System, (2013).
[2] D. C. Montgomery, JohnWiley and Sons, Design and Analysis of Experiments, New York. (6th ed.), (2005).
[3] R. A. Fisher, “The Use of Multiple Measurements in Taxonomic Problems”, Annals of Eugenics, Vol. 7, No. 7, pp. 179-188, (1936).
[4] M. Kano, S. Tanaka, S. Hasebe, I. Hashimoto and H. Ohno, “Combined
Multivariate Statistical Process Control, In proceedings of IFAC Symposium on Advanced Control of Chemical Plants”, ADCHEM 2004, pp. 303-308, (2004).
[5] T. T. Shannon and J. McNames, “ICA Based Disturbance Specific Control Charts, In proceedings of the IEEE International Conference on Information Reuse and Integration”, IRI 2007, pp. 251-256, (2007).
[6] A. Hyv¨arinen, “Fast and Robust Fixed-Point Algorithms for Independent Component Analysis”. IEEE Transactions on Neural Networks, Vol. 10, No. 3, pp. 626-634, (1999).
[7] A. Hyv¨arinen, “Survey on Independent Component Analysis”, Neural Computing Surveys, Vol. 2, pp. 94-128, (1999).
[8] A. Hyv¨arinen and E. Oja , “Independent Component Analysis: Algorithms and Applications”, Neural Networks, Vol. 13, pp. 411-430, (2000).
[9] A. Hyv¨arinen, J. Karhunen and E. Oja, “Independent Component Analysis”, Wiley-Interscience. (1st ed.), (2001).
[10]Dheeraj K. Rajan, Sarah Bunston, Sanjay Misra, et al., “Dysfunctional Autogenous Hemodialysis Fistulas: Outcomes after Angioplasty - Are There Clinical Predictors of Patency?” Radiology, (2004).
[11]H. A. Mansy, S.J. Hoxie, N.H. Patel, R.H. Sandier, “Computerised analysis of auscultatory sounds associated with vascular patency of haemodialysis access”, Med. Biol. Eng. Comput., pp. 56-62, (2005).
[12]MacroM. M., Pablo V. and Bengt M., “Characterisation of Arteriovenous Fistula’s Sound Recordings Using Principal Component Analysis”, 31rd Annual International Conference of the IEEE EMBS, pp. 5661-5664, (2009).
[13]Macro M. M., Pablo V. and Bengt M., “Acoustical Detection of Venous Stenosis in Hemodialysis Patients Using Principal Component Analysis”, 32nd Annual International Conference of the IEEE EMBS, pp. 3654-3657, (2010).
[14]Macro M. M. and Bengt M., “Analysis of The Vascular Sounds of The Arteriovenous Fistula’s Anastomosis”, 33rd Annual International Conference of
the IEEE EMBS, pp. 3784-3787, (2011). [15]Pablo V., Macro M. M. and Bengt M., “Arteriovenous Fistula Stenosis Detection Using Wavelets and Support Vector Machines”, 31rd Annual International Conference of the IEEE EMBS, pp. 1298-1301, (2009).
[16]Mikkel Grama, Jens Tranholm Olesena, Hans Christian Riisa, Maiuri Selvaratnama, Michalina Urbaniaka, “Stenosis detection algorithm for screening of arteriovenous fistulae”, SEMCON conference paper 1-6, (2010).
[17]Masako Kawai and Akifumi Suzuki, “Analysis and Evaluation of Vascular Sounds in Hemodialysis”, 35th Annual International Conference of the IEEE EMBS ,Osaka, Japan, 3 - 7 July, (2013).
[18]Chen W.L.,Lin C.H., Lin T.C., Chen P.J., Kan C.D., “Stenosis Detection using Burg Method with Autoregressive Model for Hemodialysis Patients”, Journal of Medical and Biological Engineering, pp. 356-362, (2012).
[19]H. H. Yang and S. I. Amari, “Adaptive Online Learning Algorithms for Blind Separation: Maximum Entropy and Minimum Mutual Information”, Neural Computation, Vol. 9, pp. 1457-1482, (1997).
[20]E. Miller and J. Fisher, “ICA Using Spacings Estimates of Entropy”, In processings of the Furth International Workshop on Independent Component Analysis and Blind Signal Separation, pp. 1271-1295, (2003).
[21]K. Hild, D. Erdogmus and J. Principe, “Blind Source Separations Using Renyi’s Mutual Information”, IEEE Signal Processing Letters, Vol. 8, No. 6, pp. 174-176, (2001).
[22]N. Delfosse and P. Loubaton, “Adaptive Blind Separation of Independent Sources: A Deflation Approach”, Signal processing, Vol. 45, pp. 59-83, (1996).
[23]C. Jutten and J. Herault, “Blind Separation of Sources, Part I: An Adaptive Algorithm Based on Neuromimetic Architecture”, Signal Processing, Vol. 24, pp. 1-10, (1991).
[24]G. H. Hardy, J. E. Littlewood and G. Polya, “Inequalities”, Cambridge. (2nd ed.), (1952).
[25]B. Ans, J. Herault and C. Jutten, “Adaptive Neural Architectures: Detection of Primitives”, In Proceedings of COGNITIVA’s 85, pp. 593-597, (1985).
[26]簡名儀, “以血流聲音時頻域特徵參數探究透析廔管之狹窄程度”, 義守大學碩士論文, (2011).
[27]陳華緯, “動靜脈透析廔管血管音擷取裝置之改良與狹窄程度之分類”, 義守大學碩士論文, (2012).
[28]江秀月, “創新的方法應用於廔管裝設與血管狹窄的血管音研究”, 清華大學碩士論文, (2013).
[29]桑慧敏, 機率與推論統計原理, McGraw-Hill International. (第一版), (2007).
[30] 第二章盲訊號分離與方法介紹, http://waoffice.ee.kuas.edu.tw/, (2017).