簡易檢索 / 詳目顯示

回結果列表
研究生： 謝敏正
Hsieh, Ming-Jeng
論文名稱： 基於斑紋對手持式超音波使用卷積神經網路的三維定位方法
Freehand Ultrasound Three-dimensional Localization Approach Based on Speckle Pattern Using Convolutional Neural Network Models
指導教授： 王士豪
Wang, Shyh-Hau
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 資訊工程學系
Department of Computer Science and Information Engineering
論文出版年： 2020
畢業學年度： 108
語文別： 英文
論文頁數： 49
中文關鍵詞： 解斑紋相關性 、卷積神經網路 、相對定位 、徒手三維超音波定位
外文關鍵詞： speckle decorrelation, convolutional neural network, relative localization, freehand 3-D ultrasound localization
相關次數： 點閱：141  下載：1 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 在非破壞性超音波檢測中，如何在不使用輔助設備的情況下進行徒手三維定位以及重建是一門值得探討的議題。過去的研究結果三維超音波影像將二維超音波影像利用影像匹配以及解斑紋相關性達到相對定位的結果。然而利用解斑紋相關性輸出直平面距離受限於斑紋之間的相關係數。本研究目的在於使用動態追蹤演算法(KLT)進行影像匹配後，對斑紋紋理利用深度學習的訓練改善傳統解斑紋相關係對於預測垂直平面距離的準確率。其中利用非等方擴散濾波器以及Nakagami 模型參數對斑紋進行斑紋狀態分析。使用步進馬達帶動商用超音波儀器搭配12MHz 探頭掃描商用組織假體驗證定位結果。結果顯示在沒有軸向與側向位移的干擾下，解斑紋相關性與使用深度學習模型所得到的平均絕對誤差為0.073mm以及0.020mm。在同時移動三軸前提下，解斑紋相關性與使用深度學習模型所得到的平均絕對誤差為0.165mm以及0.130mm。本篇研究提供基於超音波紋理對相對定位達到徒手式三維重建的可行性。

    In non-destructive ultrasound testing, how to perform 3D localization and reconstruction without using localization equipment is a worthy of being topic discussed. The previous studies have used image registration method and speckle decorrelation to achieve relative localization from 2-D slices to 3D reconstruction. However, the output of elevational distance using speckle decorrelation is limited by the correlation coefficient between frame-to-frame. Therefore, the purpose of this study is using in-plane motion algorithm (KLT) to detect and registration. Then using deep learning model training improve the accuracy of the conventional speckle decorrelation. Simultaneously, using anisotropic diffusion filter and Nakagami distribution model analysis scatterer concentration in speckle. This study used stepper motor and commercial ultrasonic scanner with a 12MHz linear probe to scan the tissue phantom to verify the localization results. The results show that without the interference of lateral and axial motion, the mean absolute error of speckle decorrelation method and deep learning model are 0.073mm and 0.020mm. Under the premise of moving the 3 axes at the same time, the mean absolute error of speckle decorrelation and deep learning model is 0.165mm and 0.130mm. This study provides a feasible image-based localization method to achieve 3-D freehand ultrasound.

    審查資格證明書 I 摘要 III ABSTRACT IV 致謝 V CONTENT VI LIST OF FIGURES VIII CHAPTER 1 INTRODUCTION 1 1.1 Foreword 1 1.2 Research Background 2 1.2.1 Non-destructed Testing 2 1.2.2 3-D Ultrasound Imaging 2 1.2.3 Localization 3 1.3 Research works 6 1.4 Research Objectives 7 CHAPTER 2 BACKGROUND 8 2.1 Fundamentals of Ultrasound 8 2.1.1 Fundamentals of Wave Propagation 8 2.1.2 Reflection, Refraction 8 2.1.3 Attenuation 10 2.2 In-plane Motion Tracking 11 2.2.1 Normalized Cross Correlation 11 2.2.2 Optical Flow Method 11 2.2.3 Kanade-Lucas-Tomasi Tracking Method 12 2.3 Speckle Detection 14 2.3.1 Speckle in Ultrasound Image 14 2.3.2 Fully Developed Speckle and Rayleigh Distribution 14 2.3.3 The Improved by Nakagami Distribution 16 2.3.4 Compression in Ultrasound Image 17 2.4 Out-of-plane Motion Estimation 19 2.4.1 Speckle Decorrelation 19 2.4.2 Decorrelation Curve 20 2.4.3 Deep Learning Model 21 CHAPTER 3 MATERIALS AND METHODS 23 3.1 In-plane Motion Tracking Method 23 3.1.1 Normalized Cross Correlation 23 3.1.2 KLT Method Based on LK Optical Flow 23 3.2 Speckle Detection 25 3.2.1 Filtering Structural Pattern 25 3.2.2 Detection Fully Developed Speckle Using Nakagami Distribution 26 3.3 Out-of-plane motion tracking method 27 3.3.1 Decorrelation Curve 27 3.3.2 Deep Learning Model 27 3.4 Verification of Image-based Localization Method 28 CHAPTER 4 RESULTS AND DISSCUSSION 31 4.1 In-plane motion Estimation Results 31 4.2 Speckle Detection Results 34 4.2.1 Fully Developed Speckle after Compression 34 4.2.2 Filtering Structural Pattern 35 4.3 Out-of-plane Motion Estimation Results 38 4.3.1 Conventional Decorrelation Curve Model 38 4.3.2 Deep Learning Model 39 4.4 3-D Reconstruction Results 43 CHAPTER 5 CONCLUSIONS 45 5.1 Conclusion 45 5.2 Future Works 46 REFERENCES 47

    1. Shung, K.K., Diagnostic ultrasound: Imaging and blood flow measurements 2005: CRC Pr I Llc. 1-220.
    2. Workman, G.L., Ultrasonic testing. 3rd ed ed. Nondestructive testing handbook. 2007, Columbus, OH: American Society for Nondestructive Testing. 588.
    3. Non-Destructive Testing Market Size. 2020, GRAND View Research, Inc. p. 83.
    4. Gholizadeh, S., A review of non-destructive testing methods of composite materials. Procedia Structural Integrity, 2016. 1: p. 50-57.
    5. Mao, X., Y. Zhao, and T. Xiao, Review of the development of metal non-destructive testing and imaging technology. 2018. 926-929.
    6. Chen, X., et al., Reconstruction of Freehand 3D Ultrasound based on Kernel Regression. Biomedical engineering online, 2014. 13: p. 124.
    7. Octorina Dewi, D.E., et al., Position Tracking Systems for Ultrasound Imaging: A Survey, in Medical Imaging Technology: Reviews and Computational Applications, K.W. Lai and D.E. Octorina Dewi, Editors. 2015, Springer Singapore: Singapore. p. 57-89.
    8. Rahni, A.A.A. and I. Yahya. Obtaining translation from a 6-DOF MEMS IMU — an overview. in 2007 Asia-Pacific Conference on Applied Electromagnetics. 2007.
    9. Chen, J.-F., et al., Determination of scan-plane motion using speckle decorrelation: Theoretical considerations and initial test. International Journal of Imaging Systems and Technology, 1997. 8.
    10. Gao, H., et al., Wireless and sensorless 3D ultrasound imaging. Neurocomputing, 2016. 195: p. 159-171.
    11. Lang, A., et al., Multi-modal registration of speckle-tracked freehand 3D ultrasound to CT in the lumbar spine. Medical Image Analysis, 2012. 16(3): p. 675-686.
    12. Housden, R., et al., Sensorless Reconstruction of Freehand 3D Ultrasound Data. Vol. 9. 2006. 356-63.
    13. Gee, A., et al., Sensorless freehand 3D ultrasound in real tissue: Speckle decorrelation without fully developed speckle. Medical image analysis, 2006. 10: p. 137-49.
    14. Prager, R.W., et al., Sensorless freehand 3-D ultrasound using regression of the echo intensity. Ultrasound Med Biol, 2003. 29(3): p. 437-46.
    15. Tetrel, L., H. Chebrek, and C. Laporte, Learning for Graph-Based Sensorless Freehand 3D Ultrasound. 2016. 205-212.
    16. Shung, K.K., Diagnostic ultrasound: Imaging and blood flow measurements. 2015: CRC press.
    17. Lewis, J.P., Fast Normalized Cross-Correlation. Ind. Light Magic, 2001. 10.
    18. Gibson, J.J., The ecological approach to visual perception. 2015.
    19. Lucas, B. and T. Kanade, An Iterative Image Registration Technique with an Application to Stereo Vision (IJCAI). Vol. 81. 1981.
    20. Horn, B.K.P. and B.G. Schunck, Determining optical flow. Artificial Intelligence, 1981. 17(1): p. 185-203.
    21. Dosovitskiy, A., et al., FlowNet: Learning Optical Flow with Convolutional Networks. 2015. 2758-2766.
    22. Jianbo, S. and Tomasi. Good features to track. in 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. 1994.
    23. Lowe, D.G. Object recognition from local scale-invariant features. in Proceedings of the Seventh IEEE International Conference on Computer Vision. 1999.
    24. Bay, H., et al., Speeded-Up Robust Features (SURF). Computer Vision and Image Understanding, 2008. 110(3): p. 346-359.
    25. Rosten, E. and T. Drummond. Machine Learning for High-Speed Corner Detection. in Computer Vision – ECCV 2006. 2006. Berlin, Heidelberg: Springer Berlin Heidelberg.
    26. Anderson, M.E., M.S. McKeag, and G.E. Trahey, The impact of sound speed errors on medical ultrasound imaging. J Acoust Soc Am, 2000. 107(6): p. 3540-8.
    27. Dutt, V. and J.F. Greenleaf, Ultrasound Echo Envelope Analysis Using a Homodyned K Distribution Signal Model. Ultrasonic Imaging, 1994. 16(4): p. 265-287.
    28. Molthen, R.C., et al., Comparisons of the Rayleigh and K-distribution models using in vivo breast and liver tissue. Ultrasound in Medicine & Biology, 1998. 24(1): p. 93-100.
    29. Dumane, V.A., et al., Computer aided classification of masses in ultrasonic mammography. Medical Physics, 2002. 29(9): p. 1968-1973.
    30. Papoulis, A. and U. Pillai, Probability, Random Variables, and Stochastic Processes, Fourth Edition. 2002.
    31. Shankar, P.M., et al., Classification of ultrasonic B-mode images of breast masses using Nakagami distribution. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2001. 48(2): p. 569-580.
    32. Narayanan, V.M., P.M. Shankar, and J.M. Reid, Non-Rayleigh statistics of ultrasonic backscattered signals. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 1994. 41(6): p. 845-852.
    33. Burckhardt, C.B., Speckle in ultrasound B-mode scans. IEEE Transactions on Sonics and Ultrasonics, 1978. 25(1): p. 1-6.
    34. Michael, F.I., et al., Analysis Of Ultrasound Image Texture Via Generalized Rician Statistics. Optical Engineering, 1986. 25(6): p. 743-748.
    35. Tuthill, T.A., R.H. Sperry, and K.J. Parker, Deviations from Rayleigh statistics in ultrasonic speckle. Ultrasonic Imaging, 1988. 10(2): p. 81-89.
    36. Wagner, R.F., M.F. Insana, and D.G. Brown, Statistical properties of radio-frequency and envelope-detected signals with applications to medical ultrasound. Journal of the Optical Society of America. A, Optics and image science, 1987. 4(5): p. 910-922.
    37. Shankar, p.m., A general statistical model for ultrasonic scattering from tissues. IEEE transactions on ultrasonics, ferroelectrics, and frequency control, 2000. 47: p. 727-36.
    38. Tsui, P.-H., C.-C. Huang, and S.-H. Wang, Use of Nakagami distribution and logarithmic compression in ultrasonic tissue characterization. Journal of Medical and Biological Engineering, 2006. 26: p. 69-73.
    39. Lin, Y.-H., et al., Assessment of the Kinetic Trajectory of the Median Nerve in the Wrist by High-Frequency Ultrasound. Sensors, 2014. 14(5): p. 7738-7752.
    40. Liu, F. and J. Liu, Anisotropic diffusion for image denoising based on diffusion tensors. Journal of Visual Communication and Image Representation, 2012. 23: p. 516–521.
    41. Perona, P. and J. Malik, Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990. 12(7): p. 629-639.
    42. Yongjian, Y. and S.T. Acton, Speckle reducing anisotropic diffusion. IEEE Transactions on Image Processing, 2002. 11(11): p. 1260-1270.
    43. Xiaona, Z. and W. Tianfu. An anisotropic diffusion filter for ultrasonic speckle reduction. in 2008 5th International Conference on Visual Information Engineering (VIE 2008). 2008.
    44. Prevost, R., et al., 3D Freehand Ultrasound Without External Tracking Using Deep Learning. Medical Image Analysis, 2018. 48.
    45. Wang, Z., et al., Image Quality Assessment: From Error Visibility to Structural Similarity. Image Processing, IEEE Transactions on, 2004. 13: p. 600-612.
    46. Krizhevsky, A., I. Sutskever, and G. Hinton, ImageNet Classification with Deep Convolutional Neural Networks. Neural Information Processing Systems, 2012. 25.
    47. Simonyan, K. and A. Zisserman, Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv 1409.1556, 2014.
    48. Chollet, F., Xception: Deep Learning with Depthwise Separable Convolutions. 2017. 1800-1807.
    49. Lin, M., Q. Chen, and S. Yan, Network In Network. 2013.

    下載圖示 校內：2025-08-31公開
    校外：2025-09-01公開
    QR CODE