臺灣博碩士論文加值系統

核磁共振影像技術(Magnetic Resonance Imaging)，由於具備了非侵入性與無輻射傷害的優點，配合強大的成像序列變化，使其提供了豐富的組織訊息，成為近年來臨床診斷及研究工作不可或缺的利器。然而，核磁共振影像在掃描影像時，由於磁場梯度不停改變，導致線圈內感應電流的產生，進而造成強大的磁力，致使線圈受力而相互碰撞，發出吵雜且令人覺得不舒服的噪音，而這噪音伴隨著高速成像序列的變化，更有明顯的激增。文獻指出，長期處於核磁共振儀之噪音下，會產生暫時性聽覺障礙。目前大部分醫院多是使用耳塞或是耳機等被動式的方法來解決此問題，然而，耳塞、耳機對低頻噪音的抑制效果非常差，也同時造成病患與操作員間溝通上的困難。因此，本研究嘗試以主動式噪音抑制的方法來克服此問題。
傳統上，主動式噪音抑制系統大多採用適應性濾波器（adaptive filter）來實現，不過仍有其缺點，如FIR濾波器不適合於寬頻的系統中使用，且其硬體電路較為龐大，而IIR濾波器於參數調整中，容易出現系統不穩定的情況。因此，我們採用具有高精確預測能力的神經網路方法，對不同的脈衝序列（pulse sequence）與掃描參數的噪音進行噪音抑制。
抑制的模擬結果顯示，對臨床上經常使用的gradient echo、spin echo及RARE-IR 成像序列，以及大腦功能性研究經常使用的echo-planar imaging成像序列進行噪音抑制，得到噪音抑制量皆在18分貝左右，而echo-planar imaging更有21分貝的噪音抑制量。在噪音抑制系統的實現上，我們架構了一個以DSP為基礎之系統，但受限於硬體系統的輸入輸出時間太長，只得到5.035分貝的噪音抑制量，若未來能提昇此系統的硬體速度，將可克服時間延遲問題，達到更好的噪音抑制結果。
關鍵字：主動式噪音抑制、核磁共振影像

Magnetic Resonance Imaging (MRI) is one of the fast growing technology in medical imaging. With its non-radioactive, and non- invasive properties, MRI has contributed to a tremendous advance in medical diagnosis technology. However, it still has some limitations and problems. The loud acoustic noise generated by gradient switching is annoying patients. According to researches, noise generating by clinical MRI would up to 100dB. When using the rapid imaging method- echo-planar imaging (EPI), it will have more serious acoustic problems, because of higher slew rate and duty cycle.
A simpler and perhaps more widely used technique is the use of ear plugs but this method seems to protect only against sounds transmitted by the auditory canal to the ear. The use of passive ear protection is not a complete solution. In general, high frequency components are adequately suppressed by around 20 dB by ear protection. But it doesn’t have such a good effect at low frequencies. Thus, we use active noise cancellation method to solve this problem.
The results of simulation show that the average attenuation level is about 18.23dB in the pulse sequences of spin echo, about 17.82dB in the pulse sequences of gradient echo, and about 20.73dB in the pulse sequences of EPI. In hardware implementation, we constructed a DSP system for active noise cancellation to cancel the noise of EPI in real-time. Due to the speed limitation of current system, the noise is attenuated at the noise level of 5.035 dB. In the future, with higher speech CPU or DSP I/O capability, we are able to solve the time-delay problem for much better noise reduction.
Keyword：Active Noise Cancellation、Magnetic Resonance Imaging

中文摘要
英文摘要
第一章 緒 論
1－1 研究動機……………………………………..…………….1
1－2 論文架構……………………………………..…………….3
第二章 相關文獻回顧
2－1 MRI噪音的來源與產生原因……………….…………….5
2－2 MRI噪音量之分析………………………….…………….8
2－3 MRI噪音對人體產生的影響…………….………………14
2－4 對抗MRI噪音的方法……………………………………17
2－5 MRI噪音對大腦功能性磁振造影的影響………………22
第三章 原 理
3－1 主動式噪音抑制的發展歷程…………………………….24
3－2 主動式噪音抑制系統的結構選擇……………………….26
3-2-1 前饋型主動式噪音抑制系統.…………………….26
3-2-2 回饋型主動式噪音抑制系統…….……………….28
3－3 類神經網路於主動式噪音抑制………………………….29
3-3-1 類神經網路之原理……………………………….30
3-3-2 類神經網路之運作……………………………….34
3-3-3 類神經網路參數對系統效能的影響………….…37
3－4 實驗架構………………………………………………….39
第四章 噪音抑制之模擬結果
4－1 類神經網路參數的最佳化調整………………….………40
4-1-1 噪音抑制量的算法………….……………………40
4-1-2 不同的神經網路處理單元數對噪音抑制的影響...40
4-1-3 最大循環次數（Max Cycle）……………………..42
4-1-4 學習速率（learning rate）…………..…………….44
4-1-5 訓練範例大小對系統效能的影響..…………….…45
4-1-6 輸入資料大小對噪音抑制的影響…………..…….46
4－2 使用者介面與步驟…………………………………….48
4-2-1 使用者介面……………………..………………….48
4-2-2 系統使用步驟………………………………...……49
4－3 MRI的噪音抑制模擬結果………….……...………….50
4-3-1 對GE噪音的模擬結果……….………………..….50
4-3-2 對EPI噪音的模擬結果……………….……...……52
4-3-3 對其他脈衝序列的模擬結果……...………………54
4－4 語音的噪音抑制模擬結果……………………………….57
4-4-1 信號雜訊比（SNR）的計算方法………………….57
4-4-2 模擬結果……...…………………………………....58
第五章 硬體系統之實現
5－1 即時噪音抑制系統的架構說明……………………….61
5－2 時間延遲問題………………………………………….62
5－3 DSP硬體實現架構………………………………….…66
第六章 討論、結論與未來展望
6－1 討論…………………………………………………….71
6－2 結論………………………………………………….…75
6－3 未來展望…………………………………………….…76
參考資料……………………………………………………………………...77

參 考 資 料
[1]、R.E.Brummett, J. Micheal Talbot, Peter Charuhas, 1988, “Potential Hearing Loss Resulting from MR Imaging.” Radiology, vol.169, pp.539-540.
[2]、Mark E. Quirk, Audrew J. Letendre, Rogert A Ciottone, James F Lingley, 1989, ”Anxiety in Patients Undergoing MR Imaging.” Radiology, Vol.170, pp.463-466.
[3]、C.T.W. Moonen, P.A. Bandettini（Editor）, 2000, Functional MRI, Springer Verlag.
[4]、David L. Pricee, “Investigation of Acoustic Noise on 15 MRI Scanners from 0.2T to 3T.” 2001, Journal of magnetic Resonance Imaging, 13: pp288-293.
[5]、Frank G. Shellock, Stacy M. Morisoli, Mokhtar Ziarati, 1994, “Measurement of Acoustic Noise during MR Imaging：Evaluation of Six Worst-Case Pulse Sequences.” Radiology, 191: pp91-93.
[6]、John R. Foster, “Sound-Level Measurements and Calculations of Safe Noise Dosage During EPI At 3T.” 2000, Journal of Magnetic Resonance Imaging, 12:157-163.
[7]、Mark McJury, Frank G. Shekkock, ”Auditory Noise Associated With MR Procedures：A Review.” 2000, Journal of Magnetic Resonance Imaging, 12:pp37-45.
[8]、M.J. McJury, 1994, “Acoustic Noise Levels Generated during High Field MR Imaging.” Clinical Radiology, 50: pp331-334.
[9]、Tosiaki Miyati ..etc , 1999, “Acoustic Noise Analysis in Echo Planar Imaging： Multicenter Trial and Comparison with Other Pulse Sequence.” IEEE Transaction on Medical Imaging, Vol.18 No.8, pp.733-736.
[10]、M McJury, A Blug, C Joerger, B Condon, D Wyper, 1994, “Short Communication：Acoustic Noise Levels during MR Imaging Scanning at 1.5T.” The British Journal of Radiology, 67: pp413-415.
[11]、Peter Mansfield, Barrry L.W. Chapman, Richard Bowtell, Paul Glover, Ron Coxon, 1995, “Active Acoustic Screening：Reduction of Noise in Gradient Coils by Lorentz Force Balancing.” MRM, 33: pp276-281.
[12]、Robert A. Hedeen, William A. Edelstein, 1997, “Characterization and Prediction of Gradient Acoustic Noise in MR Imaging.” MRM, 37: pp7-10.
[13]、Robert Hurwitz, Samuel R. Lane Robert A. Bell, Michael N. Brant-Zawadzki, 1989, “Acoustic Analysis of Gradient-Coil Noise in MR Imaging.” Radiology, 173: pp545-548.
[14]、Z.H Cho, S.H. Park, J.H. Kim, S.C. Chung, S.T. Chung, C.W. Moon, J.H. Yi, C.H. Sin, E.K. wong, 1997, “Analysis of Acoustic Noise in MRI.” Magnetic Resonance Imaging, Vol.15 No.7.
[15]、Arnold M. Goldman, William E. Gossman, Paul C. Friedlander, 1989, “Reduction of Sound Level with Anti-noise in MR Imaging.” Radiology,173: pp.549-550.
[16]、Casper K. Chen, Jyh-Horng Chen, Tiz-Dar Chiueh, 1999, “Active Cancellation System of Acoustic Noise in MR Imaging.” IEEE Transaction on Biomedical Engineering, Vol.46 No 2, pp.186-191.
[17]、M McJury, R.W. Stewart, D. Crawford, E. Toma, 1997, “The Use of Active Noise Control to Reduce Acoustic Noise Generated During MRI Scanning：Some Initial Results.” Magnetic Resonance Imaging, Vol.15 No.3, pp.319-322.
[18]、林俊成，「主動式噪音消除技術之研究」，碩士論文，台灣科技大學電機工程技術研究所，民國85年。
[19]、黃世昌，「主動式噪音控制技術於風扇與壓縮機噪音降低之應用」，碩士論文，交通大學控制工程研究所，民國85年。
[20]、Casper K. Chen, 1998, “Active Noise Control Using Neural Network.” Ph.D thesis, National Taiwan University, Department of EE.
[21]、馬仲宏，「使用定點數位訊號處理器實現適應性單通道噪音抑制系統」，碩士論文，成功大學電機工程研究所，民國87年。
[22]、葉怡成，「類神經網路模式應用與實作」，儒林書局，民國84年第四版。
[23]、Allen Counter, Ake Olofsson, H.F. Grahn, Erik Bong, 1997, “MRI Acoustic Noise：Sound Pressure and Frequency Analysis.” JMRI, 7: pp606-611.
[24]、Frank G. Shllock, Mokhtar Ziarati, Dennis Atkinson, Dar-Yeong Chen, 1998, “Determination of Gradient Magnetic Field-Indurced Acoustic Noise Associated with the Use of Echo Planar and Three Dimensional, Fast Spin Echo Techniques.” JMRI, 8: pp1154-1157.