帳號:guest(3.148.115.173)          離開系統
字體大小: 字級放大   字級縮小   預設字形  

詳目顯示

以作者查詢圖書館館藏以作者查詢臺灣博碩士論文系統以作者查詢全國書目
作者(中文):李威儒
作者(外文):Lee, Wei-Ru.
論文名稱(中文):基於深度學習之神經細胞軸突影像切割
論文名稱(外文):Deep Learning for Neuron Dendrite Segmentation
指導教授(中文):吳順吉
指導教授(外文):Wu, Shun-Chi
口試委員(中文):王翊青
張順福
口試委員(外文):Wang, I-Ching
Chang, Shun-Fu
學位類別:碩士
校院名稱:國立清華大學
系所名稱:工程與系統科學系
學號:106011551
出版年(民國):109
畢業學年度:108
語文別:中文
論文頁數:41
中文關鍵詞:深度學習影像切割
外文關鍵詞:deep learningimage segmentation
相關次數:
  • 推薦推薦:0
  • 點閱點閱:211
  • 評分評分:*****
  • 下載下載:0
  • 收藏收藏:0
神經細胞軸突的分化程度在醫療以及生物醫學上是重要的研究議題,例如:為了測試材料對刺激神經細胞分化的好壞,將細胞株養殖在材料上並且拍攝分化的影像,藉由計算分化後神經細胞軸突的長度,即可得知該材料的效果。雖然現今影像獲取的技術已經相當進步,但後續的影像分析上,以人工的方式處理會需要研究員自身長期的經驗累積來判斷,消耗過多人力資源,故而自動化影像切割方式的引入可以協助解決此問題。近年來,隨著神經網路的發展,運用在醫學上的影像切割及影像辨識的案例已經相當常見,也有著很好的成果,再加上圖形處理器(Graphics Processing Unit, GPU)的使用可以增加神經網路的訓練,因此運用深度學習(deep learning)來進行此課題相關的影像處理可以是一個選項。
本論文提供了兩種自動化的影像切割的方式來做神經細胞軸突的切割,分別為基於影像處理(image processing)技術的影像切割以及基於深度學習的影像切割方法。基於影像處理技術的影像切割方式是組合許多影像處理的方式將一張複雜的細胞分化影像中切割出神經軸突;而基於深度學習的影像切割方式則是將影像中神經細胞軸突與其他物件做標記區分後來訓練分類模型,再輸入影像做細胞軸突的預測。
最後,我們設計了一個基於影像形態學的邊界條件消除細胞本體干擾並且將切出的細胞軸突做骨架化的修正,可以使得切出的結果更加精確,同時自動計算出其長度。基於深度學習的影像切割方式會比影像處理技術切割來的還要細緻。若輸入的影像較為複雜,基於深度學習的影像切割也具有更大的彈性可以將目標切出。
The degree of differentiation of neuron dendrite is an important research topic in medicine and biomedicine. To test the material is suitable for stimulating nerve cell differentiation, culturing the cell strain on the materials, and calculating the length of cell dendrite is essential. Although image acquisition technology has been quite advanced, it still requires much time to do subsequent image analysis. Thus, the introduction of automatic image segmentation is crucial to help solve this issue. In recent years, with the development of the neural network, image segmentation, and recognition used in medicine have been quite common, and have achieved excellent results. If the graphics processing units are used, the speed of computing in image processing can be further increased. This study provides two automatic image segmentation methods for nerve cell dendrite segmentation: image processing technology-based and deep learning image-based segmentation methods. The image processing technology-based method is a combination of many image processing elements to realize cutting out dendrites in an intricate cell differentiation image. On the other hand, the deep learning image-based method uses the images of labeled nerve cell dendrites and other objects to train the classification model that can quickly cut out the parts of the dendrite. Finally, a morphology-based method is designed to eliminate the interference of the cell body, and the skeletonized correction of the cutting cell dendrite can make the segmentation result more accurate and automatically calculate its length.
摘要 i
Abstract ii
目錄 iv
圖目錄 vi
表目錄 viii
第1章 緒論 1
1.1 神經細胞的功用 1
1.2 研究動機 1
1.3 文獻回顧 2
1.4 研究架構 4
第2章 研究數據與前處理 5
2.1 數據來源 5
2.1.1 影像概念 5
2.2 數據前處理與增量方式 7
2.2.1 影像處理技術的切割前處理 7
2.2.2 深度學習切割的前處理及數據增量 9
第3章 影像處理技術的切割方式 10
3.1 影像切割 10
3.2 影像邊界探測 11
3.3 形態學運用 13
3.3.1 侵蝕 14
3.3.2 膨脹 15
3.3.3 影像開運算 17
第4章 深度學習切割 19
4.1 卷積神經網絡 19
4.2 InputCascadeCNN模型 21
4.3 切割影像的後處理 24
4.3.1 影像物件骨架化 24
4.3.2 消除細胞本體的殘留干擾 26
第5章 結果與討論 28
5.1 結果呈現 28
5.1.1 評量指標 28
5.1.2 影像處理技術的切割方法結果 29
5.1.3 深度學習影像切割方法結果 31
5.2 長度計算 34
5.3 比較與討論 36
第6章 總結 38
參考資料 39

[1] W. J. B. j. Rall, "Electrophysiology of a dendritic neuron model," vol. 2, no. 2, pp. 145-167, 1962.
[2] D. Johnston, J. C. Magee, C. M. Colbert, and B. R. J. A. r. o. n. Christie, "Active properties of neuronal dendrites," vol. 19, no. 1, pp. 165-186, 1996.
[3] N. J. N. R. N. Spruston, "Pyramidal neurons: dendritic structure and synaptic integration," vol. 9, no. 3, pp. 206-221, 2008.
[4] A. W. Moore, L. Y. Jan, and Y. N. J. S. Jan, "Hamlet, a binary genetic switch between single-and multiple-dendrite neuron morphology," vol. 297, no. 5585, pp. 1355-1358, 2002.
[5] Z. F. Mainen and T. J. J. N. Sejnowski, "Influence of dendritic structure on firing pattern in model neocortical neurons," vol. 382, no. 6589, pp. 363-366, 1996.
[6] O. Gökçe, "Channelrhodopsin assisted synapse identity mapping reveals clustering of layer 5 intralaminar inputs," Ludwig-Maximilians-Universität München, 2013.
[7] S. F. Sorrells et al., "Human hippocampal neurogenesis drops sharply in children to undetectable levels in adults," vol. 555, no. 7696, pp. 377-381, 2018.
[8] M. Boldrini et al., "Human hippocampal neurogenesis persists throughout aging," vol. 22, no. 4, pp. 589-599. e5, 2018.
[9] W.-H. Chan and Y.-J. J. T. l. Chang, "Dosage effects of resveratrol on ethanol-induced cell death in the human K562 cell line," vol. 161, no. 1, pp. 1-9, 2006.
[10] M. T. Knopp, "Analysis of spine plasticity in CA1 hippocampal pyramidal neurons employing live cell nanoscopic imaging," Ludwig-Maximilians-Universität München, 2014.
[11] K. A. Al-Kofahi et al., "Rapid automated three-dimensional tracing of neurons from confocal image stacks," vol. 6, no. 2, pp. 171-187, 2002.
[12] Z. Vasilkoski and A. J. J. o. n. m. Stepanyants, "Detection of the optimal neuron traces in confocal microscopy images," vol. 178, no. 1, pp. 197-204, 2009.
[13] P. Chothani, V. Mehta, and A. J. N. Stepanyants, "Automated tracing of neurites from light microscopy stacks of images," vol. 9, no. 2-3, pp. 263-278, 2011.
[14] M. Nacucchi, F. De Pascalis, M. Scatto, L. Capodieci, and R. Albertoni, "Structural analysis of advanced polymeric foams by means of high resolution X-ray computed tomography," in AIP Conference Proceedings, 2016, vol. 1749, no. 1, p. 020009: AIP Publishing LLC.
[15] N. Karmakar, S. Mondal, and A. J. I. S. Biswas, "Determination of 3D curve skeleton of a digital object," vol. 499, pp. 84-101, 2019.
[16] R. C. Gonzalez, R. E. Woods, and S. L. Eddins, Digital image processing using MATLAB. Pearson Education India, 2004.
[17] A. Perez, R. C. J. I. t. o. p. a. Gonzalez, and m. intelligence, "An iterative thresholding algorithm for image segmentation," no. 6, pp. 742-751, 1987.
[18] B. Gayathri and P. Raajan, "A survey of breast cancer detection based on image segmentation techniques," in 2016 International Conference on Computing Technologies and Intelligent Data Engineering (ICCTIDE'16), 2016, pp. 1-5: IEEE.
[19] Y. J. J. P. r. Zhang, "A survey on evaluation methods for image segmentation," vol. 29, no. 8, pp. 1335-1346, 1996.
[20] W. S. McCulloch and W. J. T. b. o. m. b. Pitts, "A logical calculus of the ideas immanent in nervous activity," vol. 5, no. 4, pp. 115-133, 1943.
[21] A. G. Ivakhnenko and V. G. Lapa, Cybernetics and forecasting techniques. New York: American Elsevier Pub. Co., 1967.
[22] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
[23] O. Ronneberger, P. Fischer, and T. Brox, "U-net: Convolutional networks for biomedical image segmentation," in International Conference on Medical image computing and computer-assisted intervention, 2015, pp. 234-241: Springer.
[24] A. Khan, A. Sohail, U. Zahoora, and A. S. J. A. I. R. Qureshi, "A survey of the recent architectures of deep convolutional neural networks," pp. 1-62, 2020.
[25] M. J. a. p. a. Wang, "Multi-path convolutional neural networks for complex image classification," 2015.
[26] K. Song, X. Tan, D. He, J. Lu, T. Qin, and T.-Y. J. a. p. a. Liu, "Double path networks for sequence to sequence learning," 2018.
[27] S. Sedlar, "Brain Tumor Segmentation Using a Multi-path CNN Based Method," 2018, pp. 403-422.
[28] J. Pawley, Handbook of biological confocal microscopy. Springer Science & Business Media, 2006.
[29] R. C. Gonzalez and R. E. Woods, Digital Image Processing. Addison-Wesley Longman Publishing Co., Inc., 2001.
[30] N. Tanaka, "Method of smoothing image signals," ed: Google Patents, 1989.
[31] P. F. Felzenszwalb and D. P. J. I. j. o. c. v. Huttenlocher, "Efficient graph-based image segmentation," vol. 59, no. 2, pp. 167-181, 2004.
[32] E. Nadernejad, S. Sharifzadeh, and H. J. A. M. S. Hassanpour, "Edge detection techniques: evaluations and comparisons," vol. 2, no. 31, pp. 1507-1520, 2008.
[33] W. Gao, X. Zhang, L. Yang, and H. Liu, "An improved Sobel edge detection," in 2010 3rd International conference on computer science and information technology, 2010, vol. 5, pp. 67-71: IEEE.
[34] O. R. Vincent and O. Folorunso, "A descriptive algorithm for sobel image edge detection," in Proceedings of Informing Science & IT Education Conference (InSITE), 2009, vol. 40, pp. 97-107: Informing Science Institute California.
[35] H. J. J. S. r. Heijmans, "Mathematical morphology: A modern approach in image processing based on algebra and geometry," vol. 37, no. 1, pp. 1-36, 1995.
[36] H. Guo, N. Ono, and S. Sagayama, "A structure-synthesis image inpainting algorithm based on morphological erosion operation," in 2008 Congress on Image and Signal Processing, 2008, vol. 3, pp. 530-535: IEEE.
[37] E. Codaro et al., "An image processing method for morphology characterization and pitting corrosion evaluation," vol. 334, no. 1-2, pp. 298-306, 2002.
[38] P. T. Jackway, M. J. I. t. o. p. a. Deriche, and m. intelligence, "Scale-space properties of the multiscale morphological dilation-erosion," vol. 18, no. 1, pp. 38-51, 1996.
[39] P. Maragos, R. J. I. T. o. A. Schafer, Speech,, and S. Processing, "Morphological skeleton representation and coding of binary images," vol. 34, no. 5, pp. 1228-1244, 1986.
[40] Y. LeCun, L. Bottou, Y. Bengio, and P. J. P. o. t. I. Haffner, "Gradient-based learning applied to document recognition," vol. 86, no. 11, pp. 2278-2324, 1998.
[41] M. Havaei et al., "Brain tumor segmentation with deep neural networks," vol. 35, pp. 18-31, 2017.
[42] X. Xiao, M. Djurisic, A. Hoogi, R. W. Sapp, C. J. Shatz, and D. L. J. J. o. n. m. Rubin, "Automated dendritic spine detection using convolutional neural networks on maximum intensity projected microscopic volumes," vol. 309, pp. 25-34, 2018.

 
 
 
 
第一頁 上一頁 下一頁 最後一頁 top
* *