請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/84505
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 郭柏齡(Po-Ling Kuo) | |
dc.contributor.author | Yi-Ting Hung | en |
dc.contributor.author | 洪怡庭 | zh_TW |
dc.date.accessioned | 2023-03-19T22:13:45Z | - |
dc.date.copyright | 2022-10-07 | |
dc.date.issued | 2022 | |
dc.date.submitted | 2022-09-23 | |
dc.identifier.citation | [1] B. Bilecenoglu, A. Uz, and N. Karalezli, “Possible anatomic structures causing entrapment neuropathies of the median nerve: an anatomic study,” Acta Orthop. Belg., vol. 71, no. 2, pp. 169–176, Apr. 2005. [2] Luchetti R, Schoenhuber R, Alfarano M, Deluca S, de Cicco G, and Landi A, “Carpal tunnel syndrome: correlations between pressure measurement and intraoperative electrophysiological nerve study.,” Muscle Nerve. , vol. 13, no. 12, pp. 1164–1168, 1990, doi: 10.1002/mus.880131211. [3] J. Becker et al., “An evaluation of gender, obesity, age and diabetes mellitus as risk factors for carpal tunnel syndrome,” Clinical Neurophysiology, vol. 113, no. 9, pp. 1429–1434, 2002, doi: https://doi.org/10.1016/S1388-2457(02)00201-8. [4] I. Atroshi, C. Gummesson, R. Johnsson, E. Ornstein, J. Ranstam, and I. Rosén, “Prevalence of Carpal Tunnel Syndrome in a General Population,” JAMA, vol. 282, no. 2, pp. 153–158, Jul. 1999, doi: 10.1001/jama.282.2.153. [5] P. Yazdanpanah, S. Aramesh, A. Mousavizadeh, Z. Ghaffari P and Khosravi, and A. Khademi, “Prevalence and severity of carpal tunnel syndrome in women,” Iran. J. Public Health, vol. 41, no. 2, pp. 105–110, Feb. 2012. [6] I. Ibrahim, W. S. Khan, N. Goddard, and P. Smitham, “Carpal tunnel syndrome: a review of the recent literature,” Open Orthop. J., vol. 6, no. 1, pp. 69–76, Feb. 2012. [7] C. H. Fernandes, L. M. Meirelles, L. R. Raduan Neto Jorge and Nakachima, J. B. G. dos Santos, and F. Faloppa, “Carpal tunnel syndrome with thenar atrophy: evaluation of the pinch and grip strength in patients undergoing surgical treatment,” Hand (N. Y.), vol. 8, no. 1, pp. 60–63, Mar. 2013. [8] A. Y. T. Lim and S. J. Sebastin, “CHAPTER 14 - Clinical examination and diagnosis,” in Practical Management of Pediatric and Adult Brachial Plexus Palsies, K. C. Chung, L. J.-S. Yang, and J. E. McGillicuddy, Eds. Philadelphia: W.B. Saunders, 2012, pp. 173–197. doi: https://doi.org/10.1016/B978-1-4377-0575-1.00014-9. [9] J. N. Katz et al., “The Carpal Tunnel Syndrome: Diagnostic Utility of the History and Physical Examination Findings,” Ann Intern Med, vol. 112, no. 5, pp. 321–327, Mar. 1990, doi: 10.7326/0003-4819-112-5-321. [10] L. Heller, H. Ring, H. Costeff, and P. Solzi, “Evaluation of Tinel’s and Phalen’s signs in diagnosis of the carpal tunnel syndrome,” Eur. Neurol., vol. 25, no. 1, pp. 40–42, 1986. [11] N. Buch-Jaeger and G. Foucher, “Correlation of clinical signs with nerve conduction tests in the diagnosis of carpal tunnel syndrome,” J. Hand Surg. Br., vol. 19, no. 6, pp. 720–724, Dec. 1994. [12] K. A. Kuhlman and W. J. Hennessey, “SENSITIVITY AND SPECIFICITY OF CARPAL TUNNEL SYNDROME SIGNS1,” Am J Phys Med Rehabil, vol. 76, no. 6, 1997, [Online]. Available: https://journals.lww.com/ajpmr/Fulltext/1997/11000/SENSITIVITY_AND_SPECIFICITY_OF_CARPAL_TUNNEL.4.aspx [13] M. R. Ghavanini and M. Haghighat, “Carpal tunnel syndrome: reappraisal of five clinical tests,” Electromyogr. Clin. Neurophysiol., vol. 38, no. 7, pp. 437–441, Oct. 1998. [14] J. C. Watson, “The electrodiagnostic approach to carpal tunnel syndrome,” Neurol. Clin., vol. 30, no. 2, pp. 457–478, May 2012. [15] I. Ibrahim, W. S. Khan, N. Goddard, and P. Smitham, “Carpal tunnel syndrome: a review of the recent literature,” Open Orthop. J., vol. 6, no. 1, pp. 69–76, Feb. 2012. [16] R. A. Werner and M. Andary, “Carpal tunnel syndrome: pathophysiology and clinical neurophysiology,” Clinical Neurophysiology, vol. 113, no. 9, pp. 1373–1381, 2002, doi: https://doi.org/10.1016/S1388-2457(02)00169-4. [17] M. Swash and M. S. Schwartz, “Nerve Entrapment and Compression Syndromes and Other Mononeuropathies,” in Neuromuscular Diseases: A Practical Approach to Diagnosis and Management, M. Swash and M. S. Schwartz, Eds. London: Springer London, 1997, pp. 133–160. doi: 10.1007/978-1-4471-3834-1_8. [18] D. F. Salerno, A. Franzblau, R. A. Werner, T. J. Bromberg M B and Armstrong, and J. W. Albers, “Median and ulnar nerve conduction studies among workers: normative values,” Muscle Nerve, vol. 21, no. 8, pp. 999–1005, Aug. 1998. [19] D. S. Stetson, B. A. Silverstein, W. M. Keyserling, R. A. Wolfe, and J. W. Albers, “Median sensory distal amplitude and latency: comparisons between nonexposed managerial/professional employees and industrial workers,” Am. J. Ind. Med., vol. 24, no. 2, pp. 175–189, Aug. 1993. [20] S. Sunderland, “The nerve lesion in the carpal tunnel syndrome.,” Journal of Neurology, Neurosurgery & Psychiatry, vol. 39, no. 7, p. 615, Jul. 1976, doi: 10.1136/jnnp.39.7.615. [21] K. R. Mills, “The basics of electromyography,” Journal of Neurology, Neurosurgery & Psychiatry, vol. 76, no. suppl 2, p. ii32, Jun. 2005, doi: 10.1136/jnnp.2005.069211. [22] Preston DC and Shapiro BE, Electromyography and Neuromuscular Disorders E-Book: Clinical-Electrophysiologic-Ultrasound Correlations: Elsevier Health Sciences. 2020. [23] R. A. Werner and M. Andary, “Electrodiagnostic evaluation of carpal tunnel syndrome,” Muscle Nerve, vol. 44, no. 4, pp. 597–607, Oct. 2011. [24] C. K. Jablecki et al., “Practice parameter: Electrodiagnostic studies in carpal tunnel syndrome [RETIRED],” Neurology, vol. 58, no. 11, p. 1589, Jun. 2002, [Online]. Available: http://n.neurology.org/content/58/11/1589.abstract [25] H. R. Koyuncuoglu, S. Kutluhan, A. Yesildag, O. Oyar, K. Guler, and A. Ozden, “The value of ultrasonographic measurement in carpal tunnel syndrome in patients with negative electrodiagnostic tests,” Eur J Radiol, vol. 56, no. 3, pp. 365–369, 2005, doi: https://doi.org/10.1016/j.ejrad.2005.05.013. [26] J. C. Witt, J. G. Hentz, and J. C. Stevens, “Carpal tunnel syndrome with normal nerve conduction studies,” Muscle Nerve, vol. 29, no. 4, pp. 515–522, Apr. 2004. [27] J. R. Fowler, J. P. Gaughan, and A. M. Ilyas, “The sensitivity and specificity of ultrasound for the diagnosis of carpal tunnel syndrome: a meta-analysis,” Clin. Orthop. Relat. Res., vol. 469, no. 4, pp. 1089–1094, Apr. 2011. [28] I. Duncan, P. Sullivan, and F. Lomas, “Sonography in the diagnosis of carpal tunnel syndrome,” AJR Am. J. Roentgenol., vol. 173, no. 3, pp. 681–684, Sep. 1999. [29] E. Kurca, V. Nosal, M. Grofik, S. Sivak, M. Turcanova-Koprusakova, and P. Kucera, “Single parameter wrist ultrasonography as a first-line screening examination in suspected carpal tunnel syndrome patients,” Bratisl. Lek. Listy, vol. 109, no. 4, pp. 177–179, 2008. [30] S. C. Roll and K. Evans, “Feasibility of Using a Hand-Carried Sonographic Unit for Investigating Median Nerve Pathology,” Journal of Diagnostic Medical Sonography, vol. 25, no. 5, pp. 241–249, Sep. 2009, doi: 10.1177/8756479309345284. [31] H. R. Koyuncuoglu, S. Kutluhan, A. Yesildag, O. Oyar, K. Guler, and A. Ozden, “The value of ultrasonographic measurement in carpal tunnel syndrome in patients with negative electrodiagnostic tests,” Eur. J. Radiol., vol. 56, no. 3, pp. 365–369, Dec. 2005. [32] W. Buchberger, W. Judmaier, G. Birbamer, M. Lener, and C. Schmidauer, “Carpal tunnel syndrome: diagnosis with high-resolution sonography.,” American Journal of Roentgenology, vol. 159, no. 4, pp. 793–798, Oct. 1992, doi: 10.2214/ajr.159.4.1529845. [33] S. C. Roll, J. Case-Smith, and K. D. Evans, “Diagnostic accuracy of ultrasonography vs. electromyography in carpal tunnel syndrome: a systematic review of literature,” Ultrasound Med. Biol., vol. 37, no. 10, pp. 1539–1553, Oct. 2011. [34] K.-I. Nakamichi and S. Tachibana, “Ultrasonographic measurement of median nerve cross-sectional area in idiopathic carpal tunnel syndrome: Diagnostic accuracy,” Muscle Nerve, vol. 26, no. 6, pp. 798–803, Dec. 2002. [35] M. Lanteri, R. Ptasznik, L. Constable, and J. K. Dawborn, “Ultrasound changes in the wrist and hand in hemodialysis patients,” Clin. Nephrol., vol. 48, no. 6, pp. 375–380, Dec. 1997. [36] W. F. Brown, G. G. Ferguson, M. W. Jones, and S. K. Yates, “The location of conduction abnormalities in human entrapment neuropathies,” Can. J. Neurol. Sci., vol. 3, no. 2, pp. 111–122, May 1976. [37] W. A. Swen, J. W. Jacobs, F. E. Bussemaker, and J. W. de Waard J W and Bijlsma, “Carpal tunnel sonography by the rheumatologist versus nerve conduction study by the neurologist,” J. Rheumatol., vol. 28, no. 1, pp. 62–69, Jan. 2001. [38] L. B. Dahlin and G. Lundborg, “The neurone and its response to peripheral nerve compression,” J. Hand Surg. Br., vol. 15, no. 1, pp. 5–10, Feb. 1990. [39] M. Keberle, M. Jenett, W. Kenn, K. Reiners, R. Peter M and Haerten, and D. Hahn, “Technical advances in ultrasound and MR imaging of carpal tunnel syndrome,” Eur. Radiol., vol. 10, no. 7, pp. 1043–1050, 2000. [40] L. Sarría, T. Cabada, R. Cozcolluela, T. Martínez-Berganza, and S. García, “Carpal tunnel syndrome: usefulness of sonography,” Eur Radiol, vol. 10, no. 12, pp. 1920–1925, 2000, doi: 10.1007/s003300000502. [41] S. C. Roll, K. D. Evans, X. Li, M. Freimer, and C. M. Sommerich, “Screening for carpal tunnel syndrome using sonography,” J. Ultrasound Med., vol. 30, no. 12, pp. 1657–1667, Dec. 2011. [42] I. Keles, A. T. Karagülle Kendi, G. Aydn, S. G. Zög, and S. Orkun, “Diagnostic Precision of Ultrasonography in Patients with Carpal Tunnel Syndrome,” Am J Phys Med Rehabil, vol. 84, no. 6, 2005, [Online]. Available: https://journals.lww.com/ajpmr/Fulltext/2005/06000/Diagnostic_Precision_of_Ultrasonography_in.7.aspx [43] J.-W. H. Korstanje et al., “Ultrasonographic assessment of longitudinal median nerve and hand flexor tendon dynamics in carpal tunnel syndrome,” Muscle Nerve, vol. 45, no. 5, pp. 721–729, May 2012. [44] T.-T. Kuo, M.-R. Lee, Y.-Y. Liao, J.-P. Chen, Y.-W. Hsu, and C.-K. Yeh, “Assessment of Median Nerve Mobility by Ultrasound Dynamic Imaging for Diagnosing Carpal Tunnel Syndrome,” PLoS One, vol. 11, no. 1, pp. e0147051-, Jan. 2016, [Online]. Available: https://doi.org/10.1371/journal.pone.0147051 [45] L. Ozçakar, D. Palamar, A. B. Carl, and F. N. Aksakal, “Precision of novice sonographers concerning median nerve and Achilles tendon measurements,” Am. J. Phys. Med. Rehabil., vol. 90, no. 11, pp. 913–916, Nov. 2011. [46] Y. el Miedany, S. Ashour, S. Youssef, A. Mehanna, and F. A. Meky, “Clinical diagnosis of carpal tunnel syndrome: Old tests–new concepts,” Joint Bone Spine, vol. 75, no. 4, pp. 451–457, 2008, doi: https://doi.org/10.1016/j.jbspin.2007.09.014. [47] V. Gulshan et al., “Performance of a Deep-Learning Algorithm vs Manual Grading for Detecting Diabetic Retinopathy in India,” JAMA Ophthalmol, vol. 137, no. 9, pp. 987–993, Sep. 2019, doi: 10.1001/jamaophthalmol.2019.2004. [48] A. Esteva et al., “Dermatologist-level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115–118, 2017, doi: 10.1038/nature21056. [49] T. A. Tuan, T. B. Pham, J. Y. Kim, and J. M. R. S. Tavares, “Alzheimer’s diagnosis using deep learning in segmenting and classifying 3D brain MR images,” International Journal of Neuroscience, vol. 132, no. 7, pp. 689–698, Jul. 2022, doi: 10.1080/00207454.2020.1835900. [50] H. A. Leopold, J. Orchard, J. S. Zelek, and V. Lakshminarayanan, “PixelBNN: Augmenting the PixelCNN with Batch Normalization and the Presentation of a Fast Architecture for Retinal Vessel Segmentation,” J Imaging, vol. 5, no. 2, 2019, doi: 10.3390/jimaging5020026. [51] M. Byra, M. Galperin, L. Ojeda-Fournier Haydee and Olson, M. O’Boyle, C. Comstock, and M. Andre, “Breast mass classification in sonography with transfer learning using a deep convolutional neural network and color conversion,” Med. Phys., vol. 46, no. 2, pp. 746–755, Feb. 2019. [52] J. MacQueen, “Some methods for classification and analysis of multivariate observations,” 1967. [53] A. P. Dempster, N. M. Laird, and D. B. Rubin, “Maximum Likelihood from Incomplete Data via the EM Algorithm,” Journal of the Royal Statistical Society. Series B (Methodological), vol. 39, no. 1, pp. 1–38, 1977, [Online]. Available: http://www.jstor.org/stable/2984875 [54] R. C. Gonzalez, R. E. Woods, and S. L. Eddins, Digital Image Processing using MATLAB. Gatesmark, LLC, 2009. [55] J. Shi and J. Malik, “Normalized Cuts and Image Segmentation,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 22, no. 8, pp. 888–905, Aug. 2000, doi: 10.1109/34.868688. [56] J. Long, E. Shelhamer, and T. Darrell, “Fully Convolutional Networks for Semantic Segmentation,” Nov. 2014, doi: 10.48550/arxiv.1411.4038. [57] M.-H. Horng, C.-W. Yang, Y.-N. Sun, and T.-H. Yang, “DeepNerve: A New Convolutional Neural Network for the Localization and Segmentation of the Median Nerve in Ultrasound Image Sequences,” Ultrasound Med Biol, vol. 46, no. 9, pp. 2439–2452, 2020, doi: https://doi.org/10.1016/j.ultrasmedbio.2020.03.017. [58] 葉政樑, “應用新穎深度學習模型於動態超音波正中神經影像進行即時性影像分割,” 碩士論文, 國立台灣大學生醫電子與資訊學研究所, 2022. [59] Z. Akkus et al., “A Survey of Deep-Learning Applications in Ultrasound: Artificial Intelligence–Powered Ultrasound for Improving Clinical Workflow,” Journal of the American College of Radiology, vol. 16, no. 9, Part B, pp. 1318–1328, 2019, doi: https://doi.org/10.1016/j.jacr.2019.06.004. [60] N. Orlando et al., “Effect of dataset size, image quality, and image type on deep learning-based automatic prostate segmentation in 3D ultrasound,” Phys Med Biol, vol. 67, no. 7, p. 074002, Apr. 2022, doi: 10.1088/1361-6560/ac5a93. [61] J. Czajkowska, J. Juszczyk, L. Piejko, and M. Glenc-Ambroży, “High-Frequency Ultrasound Dataset for Deep Learning-Based Image Quality Assessment,” Sensors, vol. 22, no. 4, 2022, doi: 10.3390/s22041478. [62] L. Wu, J.-Z. Cheng, S. Li, T. Lei Baiying and Wang, and D. Ni, “FUIQA: Fetal ultrasound image quality assessment with deep convolutional networks,” IEEE Trans. Cybern., vol. 47, no. 5, pp. 1336–1349, May 2017. [63] R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” Nov. 2013, doi: 10.48550/arxiv.1311.2524. [64] R. Girshick, “Fast R-CNN,” in 2015 IEEE International Conference on Computer Vision (ICCV), 2015, pp. 1440–1448. doi: 10.1109/ICCV.2015.169. [65] S. Ren, K. He, R. Girshick, and J. Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks,” Jun. 2015, doi: 10.48550/arxiv.1506.01497. [66] J. R. R. Uijlings, K. E. A. van de Sande, T. Gevers, and A. W. M. Smeulders, “Selective Search for Object Recognition,” Int J Comput Vis, vol. 104, no. 2, pp. 154–171, 2013, doi: 10.1007/s11263-013-0620-5. [67] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” Jun. 2015, doi: 10.48550/arxiv.1506.02640. [68] W. Liu et al., “SSD: Single Shot MultiBox Detector,” Dec. 2015, doi: 10.1007/978-3-319-46448-0_2. [69] K. Wada, “Labelme: Image Polygonal Annotation with Python [Computer software],” 2016, doi: 10.5281/zenodo.5711226. [70] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” CoRR, vol. abs/2004.10934, 2020, [Online]. Available: https://arxiv.org/abs/2004.10934 [71] H. S. Kim, S. H. Joo, H. K. Cho, and Y. W. Kim, “Comparison of Proximal and Distal Cross-Sectional Areas of the Median Nerve, Carpal Tunnel, and Nerve/Tunnel Index in Subjects With Carpal Tunnel Syndrome,” Arch Phys Med Rehabil, vol. 94, no. 11, pp. 2151–2156, 2013, doi: https://doi.org/10.1016/j.apmr.2013.05.008. [72] M. Tan, R. Pang, and Q. v Le, “EfficientDet: Scalable and Efficient Object Detection,” CoRR, vol. abs/1911.09070, 2019, [Online]. Available: http://arxiv.org/abs/1911.09070 [73] T.-Y. Lin, P. Dollár, R. B. Girshick, K. He, B. Hariharan, and S. J. Belongie, “Feature Pyramid Networks for Object Detection,” CoRR, vol. abs/1612.03144, 2016, [Online]. Available: http://arxiv.org/abs/1612.03144 [74] S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path Aggregation Network for Instance Segmentation,” CoRR, vol. abs/1803.01534, 2018, [Online]. Available: http://arxiv.org/abs/1803.01534 [75] G. Jocher et al., “ultralytics/yolov5: v3.1 - Bug Fixes and Performance Improvements,” Oct. 2020, doi: 10.5281/ZENODO.4154370. [76] S. Elfwing, E. Uchibe, and K. Doya, “Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning,” CoRR, vol. abs/1702.03118, 2017, [Online]. Available: http://arxiv.org/abs/1702.03118 [77] C.-Y. Wang, H.-Y. M. Liao, I.-H. Yeh, Y.-H. Wu, P.-Y. Chen, and J.-W. Hsieh, “CSPNet: A New Backbone that can Enhance Learning Capability of CNN,” CoRR, vol. abs/1911.11929, 2019, [Online]. Available: http://arxiv.org/abs/1911.11929 [78] K. He, X. Zhang, S. Ren, and J. Sun, “Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition,” CoRR, vol. abs/1406.4729, 2014, [Online]. Available: http://arxiv.org/abs/1406.4729 [79] J. Hu, L. Shen, and G. Sun, “Squeeze-and-Excitation Networks,” CoRR, vol. abs/1709.01507, 2017, [Online]. Available: http://arxiv.org/abs/1709.01507 [80] S. Woo, J. Park, J.-Y. Lee, and I. S. Kweon, “CBAM: Convolutional Block Attention Module,” CoRR, vol. abs/1807.06521, 2018, [Online]. Available: http://arxiv.org/abs/1807.06521 [81] Y.-C. Liu et al., “Unbiased Teacher for Semi-Supervised Object Detection,” CoRR, vol. abs/2102.09480, 2021, [Online]. Available: https://arxiv.org/abs/2102.09480 [82] Rodner C, Raissis A, and Akelman E, “Carpal tunnel syndrome.,” Orthopaedic Knowledge Online Journal. American Academy of Orthopaedic Surgeons, vol. 7, no. 5, 2009. [83] T.-Y. Lin et al., “Microsoft COCO: Common Objects in Context,” in Computer Vision – ECCV 2014, 2014, pp. 740–755. | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/84505 | - |
dc.description.abstract | 腕隧道症候群是臨床上常見的神經纏套疾病,發生原因為腕隧道內的正中神經受周圍組織壓迫損傷而致。超音波具有非侵入性、病患接受度高等優勢,且近年來證實利用動態超音波可檢測出受纏套之正中神經動態形態特徵有異常表現的相關研究日益增加,使得動態超音波成為一項腕隧道症候群極具潛力的診斷工具。本實驗室先前研究利用深度學習模型進行動態超音波正中神經影像即時分割,可自動描繪神經輪廓並即時完成動態型態數據分析。然而為了獲得準確分割結果並確保後續分析數據可靠,需耗費大量人力事先從蒐集到的研究資料篩去不利於神經分割或後續分析的影像,如正中神經外觀模糊或拍攝之解剖切面不適當的資料,此外,若收案過後才發現蒐集到的資料不利於分析,通常也難以要求受測者回診並重新拍攝,因此若能開發一個平台使得獲取影像的同時即可自動且客觀的判斷影像是否適用於後續分析將會是一個有效的解決方式。因此在本篇研究中,我們即以資料蒐集當下自動完成影像條件評估為目標,提出了基於深度學習的目標檢測框架。有鑑於清晰明確的神經邊緣為獲得準確分割結果的重要條件且在整個影像序列中動態型態參數的評估需基於相同解剖平面,我們將影像中神經外觀清晰以及拍攝於腕隧道近端入口橫截面定義為理想影像須具備的兩項條件,並且以影像中同時存在舟狀骨與豆狀骨認定拍攝橫截面位於腕隧道近端入口處。對於正中神經外觀清晰度條件判定,我們根據其邊緣灰度變化來進行量化,再透過閾值設定將計算出的清晰度分數轉為高、低清晰度的二元分類。對於拍攝解剖平面的判定,為了節省人工標註骨骼標誌的成本,我們利用半監督式學習方法Unbiased Teacher僅使用少量人工標記的資料來自動識別整個數據集中的兩個標誌,該模型在測試集上獲得Precision、Recall、F1-Score分別為0.941、0.971、0.956的表現,並以此模型之預測結果作為骨骼標誌的Ground-truth。而後,我們將完成標記的數據用於訓練及評估端到端的自動影像條件評估模型,此處我們使用較輕量的物件偵測模型EfficientDet-D0、YOLOv5n以及在YOLOv5n的空間金字塔池化層添加注意力機制以增強特徵提取能力的三種改良架構YOLOv5n-SPPSE、YOLOv5n-SPPCBAM、YOLOv5n-SPPmultiCBAM。實驗結果顯示注意力機制的添加成功在維持即時預測的同時再提高了檢測準確度,EfficientDet-D0、YOLOv5n、YOLOv5n-SPPSE、YOLOv5n-SPPCBAM 和 YOLOv5n-SPPmultiCBAM的平均F1-Score分別為0.823、0.868、0.876、0.873及0.883,推論速度(FPS)則分別為23.02、44.78、44.40、42.96、40.71。相較於使用未經影像條件篩選前的資料集,以篩選後的數據集訓練影像分割模型YOLACT在Average IoU及Dice Coefficient各提升了6%及4.5%。綜上所述,我們提出了一個端到端的架構得以即時自動評估影像條件是否適用於神經自動分割及後續資料分析,可望成為使用動態超音波診斷腕隧道症候群時的有效輔助工具。 | zh_TW |
dc.description.abstract | Carpal tunnel syndrome (CTS) is a common entrapment neuropathy caused by compression of the median nerve by adjacent tissues as it travels through the carpal tunnel. Ultrasonography has the advantages of non-invasiveness and high patient acceptance. An increasing line of evidence indicates that an entrapped median nerve exhibits abnormal morphological dynamics as evaluated by dynamic ultrasonography. This renders dynamic ultrasonography as a promising diagnostic tool for CTS. Our laboratory has previously shown real-time segmentation of median nerve in dynamic ultrasonography using deep-learning models to automatically delineate the nerve contour and analyze the morphological dynamics as the images were acquired. To achieve high segmentation accuracy and ensure the reliability of subsequent analysis, however, it demanded tons of human labor to manually screen out images in conditions unfavorable for nerve segmentation or subsequent analysis from the acquired images, such as images with blurred or vague nerve appearance, or taken at improper anatomical plane. Of note, it is usually difficult to ask the subject to revisit the clinic to reacquire the images if the previously acquired ones were unfavorable for image analysis. One appealing solution is to provide a platform to automatically and objectively determine whether the image is favorable for subsequent image analysis simultaneously when the image is acquired. In the present work, we proposed deep learning-based, object detection frameworks aiming to automatically evaluate the image condition immediately as the image is collected. Given that well-defined nerve boundaries are imperative for correct segmentation and the morphological dynamics should be evaluated at the same anatomical plane throughout the image sequence, we referred to a favorable image condition as an image with clear nerve appearance and taken at the plane transecting the proximal inlet of carpal tunnel, defined by the coexistence of the scaphoid and the pisiform on the image. We quantified the nerve clarity according to the gray-level variation across the nerve boundaries, and converted the calculated numbers to the category of high or low clarity using thresholds. To save the human work for annotation of the bony landmarks, we trained Unbiased Teacher, a semi-supervised learning framework, using a small amount of manually labeled data to automatically recognize the two landmarks in the whole dataset. The model achieved a precision of 0.941, a recall of 0.971, and an F1-score of 0.956 in the test dataset. The predicted results were then utilized as the ground-truth for the existence of the bony landmarks. The data labelled for the favorability of the image condition were then utilized for the training and performance evaluation of the end-to-end, automatic image condition assessment frameworks using lightweight object detection models, consisting of EfficientDet-D0, YOLOv5n, and three architectures modified from YOLOv5n by introducing attention modules into spatial pyramid pooling layer to enhance the extraction of important features, which were named as YOLOv5n-SPPSE, YOLOv5n-SPPCBAM, and YOLOv5n-SPPmultiCBAM. Our results showed that the implementation of the attention mechanisms successfully improved the detection accuracy while maintaining real-time prediction. The achieved F1-scores were 0.823, 0.868, 0.876, 0.873, and 0.883 for EfficientDet-D0, YOLOv5n, YOLOv5n-SPPSE, YOLOv5n-SPPCBAM, and YOLOv5n-SPPmultiCBAM, respectively. The inference speeds were 23.02, 44.78, 44.40, 42.96, 40.71 for EfficientDet-D0, YOLOv5n, YOLOv5n-SPPSE, YOLOv5n-SPPCBAM, and YOLOv5n-SPPmultiCBAM, respectively. When compared with the segmentation performance achieved by YOLACT using dataset without screening for the image condition, the model using the screened dataset increased the performance by 6% and 4.5% in average IoU and Dice coefficient, respectively. Our works provide end-to-end frameworks automatically evaluating the favorability of image condition in real time for the nerve segmentation accuracy and the subsequent data analysis, which is highly potential to be an auxiliary tool for the automatic diagnosis of CTS using dynamic ultrasonography. | en |
dc.description.provenance | Made available in DSpace on 2023-03-19T22:13:45Z (GMT). No. of bitstreams: 1 U0001-2009202215132400.pdf: 6031284 bytes, checksum: 6d36aed4667303c8f53ec08c21e3f358 (MD5) Previous issue date: 2022 | en |
dc.description.tableofcontents | 誌謝 .................................................................................................................. I 中文摘要 ........................................................................................................... II ABSTRACT ....................................................................................................... IV 第一章、緒論 .................................................................................................... 1 1.1 腕隧道症候群 .............................................................................................. 1 1.1.1 腕隧道解剖構造 ....................................................................................... 1 1.1.2 疾病簡介...................................................................................................1 1.2 腕隧道症候群之診斷 ................................................................................... 2 1.2.1 理學測試診斷 .......................................................................................... 2 1.2.2 電生理檢查 ............................................................................................. 3 1.2.3 超音波影像 ............................................................................................. 4 1.3 自動化正中神經分割與應用 ......................................................................... 7 1.3.1 深度學習於醫學影像 ................................................................................ 7 1.3.2 影像分割方法 .......................................................................................... 8 1.3.3 正中神經相關研究 ................................................................................... 9 1.4 研究動機與目標 ......................................................................................... 10 第二章、文獻回顧與探討 .................................................................................. 13 2.1 深度學習與影像拍攝條件評估 ..................................................................... 13 2.2 深度學習之物件偵測方法 ........................................................................... 16 2.2.1 Two-Stage 模型..................................................................................... 16 2.2.1.1 R-CNN................................................................................................ 16 2.2.1.2 Fast R-CNN ....................................................................................... 17 2.2.1.3 Faster R-CNN .................................................................................... 18 2.2.2 One-Stage 模型 .................................................................................... 20 2.2.2.1 YOLO ................................................................................................. 20 2.2.2.2 SSD.................................................................................................... 22 第三章、研究方法 ............................................................................................. 24 3.1 研究流程架構.............................................................................................. 24 3.2 研究資料.................................................................................................... 26 3.2.1 資料蒐集..................................................................................................26 3.2.2 資料標註.................................................................................................26 3.2.2.1 正中神經標記.......................................................................................26 3.2.2.2 舟狀骨與豆狀骨標記............................................................................27 3.2.3 資料前處理與擴增...................................................................................28 3.3 理想影像拍攝條件定義............................................................................... 29 3.3.1 正中神經清晰度.......................................................................................29 3.3.2 影像拍攝截面..........................................................................................33 3.4 影像條件評估模型...................................................................................... 34 3.4.1 模型架構選擇...........................................................................................34 3.4.2 EfficientDet .......................................................................................... 35 3.4.3 YOLOv5 ................................................................................................ 38 3.4.4 YOLOv5 架構改良 ................................................................................. 41 3.4.4.1 YOLOv5-SPPSE ................................................................................ 41 3.4.4.2 YOLOv5-SPPCBAM .......................................................................... 42 3.4.4.3 YOLOv5-SPPmultiCBAM................................................................... 44 3.5 半監督式學習............................................................................................. 44 第四章、研究結果 ............................................................................................ 47 4.1 正中神經清晰度分數................................................................................... 47 4.1.1 分數計算結果...........................................................................................47 4.1.2 分數分佈 ............................................................................................... 48 4.1.3 影像經前、後處理之影響.........................................................................49 4.2 模型評估方式............................................................................................ 50 4.2.1 物件偵測模型..........................................................................................51 4.2.2 物件分割模型.........................................................................................52 4.3 半監督式學習輔助標註.............................................................................. 52 4.3.1 訓練資料集.............................................................................................52 4.3.2 模型偵測結果.........................................................................................53 4.4 影像條件評估模型..................................................................................... 54 4.4.1 訓練資料集.............................................................................................54 4.4.2 模型偵測結果.........................................................................................54 4.5 影像條件篩選............................................................................................ 56 4.5.1 影像級別................................................................................................56 4.5.2 訓練資料集............................................................................................57 4.5.3 模型分割結果.........................................................................................60 第五章、討論 ................................................................................................. 61 5.1 模型輔助標記............................................................................................ 61 5.1.1 訓練方式對模型效能的影響.....................................................................61 5.1.2 錯誤預測案例分析..................................................................................62 5.1.3 EMA 參數設定 ...................................................................................... 63 5.1.4 以半監督訓練模型輔助標記之優勢..........................................................64 5.2 影像條件評估模型.....................................................................................65 5.2.1 各模型預測結果比較..............................................................................65 5.2.2 注意力機制探討.....................................................................................68 5.3 影像條件篩選的影響..................................................................................71 5.3.1 剔除影像討論.........................................................................................71 5.3.2 影像條件評估之重要性...........................................................................72 第六章、結論 ................................................................................................. 74 第七章、未來工作 ........................................................................................... 76 7.1 理想影像拍攝條件...................................................................................... 76 7.2 正中神經清晰度判定.................................................................................. 76 7.3 標準影像拍攝截面判定............................................................................... 76 7.4 影像條件評估模型...................................................................................... 77 參考文獻 ................................................................................................... 78 | |
dc.language.iso | zh-TW | |
dc.title | 以注意力機制改良物件偵測模型進行腕隧道正中神經超音波影像拍攝條件即時評估 | zh_TW |
dc.title | Real-time ultrasound image condition assessment of median nerve in carpal tunnel using improved object detection model with attention mechanism | en |
dc.type | Thesis | |
dc.date.schoolyear | 110-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 吳沛遠(Pei-Yuan Wu),張瑞峰(Ruey-Feng Chang),吳爵宏(Chueh-Hung Wu) | |
dc.subject.keyword | 腕隧道,自動影像條件評估,深度學習,物件偵測,正中神經,超音波影像, | zh_TW |
dc.subject.keyword | carpal tunnel,automatic image condition assessment,deep learning,object detection,median nerve,ultrasonography, | en |
dc.relation.page | 86 | |
dc.identifier.doi | 10.6342/NTU202203652 | |
dc.rights.note | 同意授權(限校園內公開) | |
dc.date.accepted | 2022-09-25 | |
dc.contributor.author-college | 電機資訊學院 | zh_TW |
dc.contributor.author-dept | 生醫電子與資訊學研究所 | zh_TW |
dc.date.embargo-lift | 2024-09-30 | - |
顯示於系所單位: | 生醫電子與資訊學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-2009202215132400.pdf 目前未授權公開取用 | 5.89 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。