臺灣博碩士論文加值系統

白質高強度訊號（White Matter Hyperintensities, WMHs）是指在腦部磁振造影（magnetic resonance images, MRI）中出現的病變，通常與阿茲海默症（Alzheimer's Disease, AD）和認知衰退相關。偵測阿茲海默症相關的白質高強度訊號，對於診斷上是一個重大挑戰。本文將腦部磁振造影解釋為高光譜圖像，使得一種高光譜子像元檢測算法—約束能量最小化（Constrained Energy Minimization, CEM）能夠應用於解決混合像元和子像元級別的白質高強度訊號偵測問題。為了解決腦部白質高強度訊號在邊界附近的非線性混合偵測問題，本文還開發了一種非線性之約束能量最小化算法，稱為核化約束能量最小化（Kernel CEM, KCEM）。由於CEM是一種不考慮空間相關性的高光譜偵測技術，本文將約束能量最小化 (CEM) 擴展為包含空間濾波器的迭代約束能量最小化（Iterative CEM, ICEM），以捕捉空間信息來偵測腦部白質高強度訊號。本文結合了迭代約束能量最小化和核化約束能量最小化，衍生出一種新的白質高強度訊號檢測算法，稱為迭代核化約束能量最小化（Iterative KCEM, IKCEM），以改善迭代約束能量最小化和核化約束能量最小化在白質高強度訊號檢測中的表現。為了評估白質高強度訊號檢測的性能，本文使用相似性指數（Dice Similarity Index, DSI）和3D-ROC分析作為評估工具。為了顯示IKCEM的優越性，本文比較了兩個常用的基於統計參數映射分析（SPM）軟體包的算法：SPM-病灶增長算法（SPM-Lesion Growth Algorithm, SPM-LGA）和SPM-病灶預測算法（SPM-Lesion Prediction Algorithm, SPM-LPA）。

White matter hyperintensities (WMHs) are lesions in brain magnetic resonance images commonly linked to Alzheimer's disease (AD) and cognitive decline. Detecting WMHs associated with AD is a significant diagnostic challenge. In this dissertation we treat a set brain MR images as a set of hyperspectral images, allowing the use of the well-established hyperspectral subpixel detection algorithm, constrained energy minimization (CEM), to address the WMHs detection problem at both mixed pixel and subpixel levels. To handle nonlinear mixing near WMH boundaries, a nonlinear version of CEM, called kernel CEM (KCEM), is developed. Since traditional CEM does not take into account for spatial correlation, it is extended to iterative CEM (ICEM) by incorporating spatial filters to capture spatial information for WMHs detection. We combine ICEM and KCEM to create a new WMHs detection algorithm, iterative KCEM (IKCEM), which enhances the performance of both ICEM and KCEM in detecting WMHs. The detection performance is evaluated using the Dice similarity index (DSI) and 3D ROC analysis. To demonstrate the superiority of IKCEM, it is compared with two commonly used software packages, statistical parametric mapping (SPM)-based algorithms: the SPM-lesion growth algorithm (SPM-LGA) and the SPM-lesion prediction algorithm (SPM-LPA).

中文摘要 i
Abstract ii
Contents iii
List of Figures v
List of Tables vii
CHAPTER 1 Introduction 1
1.1 Motivations 2
1.2 Hyperspectral Imaging technology 5
1.3 Related Works to Hyperspectral Imaging Applications to MRI 6
1.4 Proposed Work 7
1.5 Contributions 7
CHAPTER 2 Background 8
2.1 Magnetic Resonance Imaging (MRI) 8
2.2 Correlation Band Expansion Process (CBEP) 10
2.3 Constrained Energy Minimization (CEM) 11
2.4 Kernel-Based Constrained Energy Minimization (KCEM) 12
2.5 Iterative K/CEM 13
2.6 Stopping Rule for Iterative K/CEM 16
2.7 The 4-way relationship among CEM, ICEM, KCEM and IKCEM 17
2.8 SPM-LGA/LPA 18
2.9 Otsu’s Method 19
2.10 Dice similarity index (DSI) 20
2.11 3-Dimensions Receiver Operating Characteristic (3D-ROC) 20
CHAPTER 3 24
Experimental Material and Methods 24
3.1Experimental Materials 24
3.2Pre-Processing the Images 25
3.3 Experimental Design 26
3.4 Ground Truth of WMHs lesions 27
3.5 Description of Detection Algorithm 28
3.6 Regions of Interest for Experiments 30
CHAPTER 4 33
Experimental Results and Discussions 33
4.1Visual analysis 33
4.2Quantitative Analysis 38
4.2.1 DSI Analysis 38
4.2.2 3D-ROC Analysis 40
4.3 Comparative Analysis of Test Methods 45
CHAPTER 5 Conclusion 49
Appendix 51
Reference 54

[1]J. C. de Groot, F.-E. de Leeuw, M. Oudkerk, A. Hofman, J. Jolles, and M. M. B. Breteler, “Cerebral White Matter Lesions and Depressive Symptoms in Elderly Adults,” Arch. Gen. Psychiatry, vol. 57, no. 11, pp. 1071–1076, Nov. 2000, doi: 10.1001/archpsyc.57.11.1071.
[2]F. E. de Leeuw et al., “Prevalence of cerebral white matter lesions in elderly people: a population based magnetic resonance imaging study. The Rotterdam Scan Study,” J. Neurol. Neurosurg. Psychiatry, vol. 70, no. 1, pp. 9–14, Jan. 2001, doi: 10.1136/jnnp.70.1.9.
[3]F. Fazekas et al., “Pathologic correlates of incidental MRI white matter signal hyperintensities,” Neurology, vol. 43, no. 9, pp. 1683–1689, Sep. 1993, doi: 10.1212/wnl.43.9.1683.
[4]Joan Jiménez‑Balado, Fabian Corlier, Christian Habeck, Yaakov Stern & Teal Eich1,“Effects of white matter hyperintensities distribution and clustering on late-life cognitive impairment,”ScientificReports(2022),doi.org/10.1038/s41598-022-06019-8.
[5]Cassandra J. Anor, Mahsa Dadar, D. Louis Collins, and M. Carmela Tartaglia“The Longitudinal Assessment of Neuropsychiatric Symptoms in Mild Cognitive Impairment and Alzheimer’s Disease and Their Association With White Matter Hyperintensities in the National Alzheimer’s Coordinating Center’s Uniform Data Set,”Biological Psychiatry: Cognitive Neuroscience and Neuroimaging January 2021; 6:70–78 www.sobp.org/BPCNNI.
[6]N. Hirono, H. Kitagaki, H. Kazui, M. Hashimoto, and E. Mori, “Impact of white matter changes on clinical manifestation of Alzheimer’s disease: A quantitative study,” Stroke, vol. 31, no. 9, pp. 2182–2188, Sep. 2000, doi: 10.1161/01.str.31.9.2182.
[7]J. Ramirez et al., “Lesion Explorer: a comprehensive segmentation and parcellation package to obtain regional volumetrics for subcortical hyperintensities and intracranial tissue,” NeuroImage, vol. 54, no. 2, pp. 963–973, Jan. 2011, doi: 10.1016/j.neuroimage.2010.09.013.
[8]J. M. Wardlaw et al., “Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration,” Lancet Neurol., vol. 12, no. 8, pp. 822–838, Aug. 2013, doi: 10.1016/S1474-4422(13)70124-8.
[9]M. Leite, L. Rittner, S. Appenzeller, H. H. Ruocco, and R. Lotufo, “Etiology-based classification of brain white matter hyperintensity on magnetic resonance imaging,” J. Med. Imaging, vol. 2, no. 1, Jan. 2015, doi: 10.1117/1.JMI.2.1.014002.
[10]L. C. Silbert, C. Nelson, D. B. Howieson, M. M. Moore, and J. A. Kaye, “Impact of white matter hyperintensity volume progression on rate of cognitive and motor decline,” Neurology, vol. 71, no. 2, pp. 108–113, Jul. 2008, doi: 10.1212/01.wnl.0000316799.86917.37.
[11]M. Yoshita, E. Fletcher, and C. DeCarli, “Current Concepts of Analysis of Cerebral White Matter Hyperintensities on Magnetic Resonance Imaging,” Top. Magn. Reson. Imaging TMRI, vol. 16, no. 6, pp. 399–407, Dec. 2005, doi: 10.1097/01.rmr.0000245456.98029.a8.
[12]S. M. Smith et al., “Accurate, robust, and automated longitudinal and cross-sectional brain change analysis,” NeuroImage, vol. 17, no. 1, pp. 479–489, Sep. 2002, doi: 10.1006/nimg.2002.1040.
[13]E. Gibson, F. Gao, S. E. Black, and N. J. Lobaugh, “Automatic segmentation of white matter hyperintensities in the elderly using FLAIR images at 3T,” J. Magn. Reson. Imaging, vol. 31, no. 6, pp. 1311–1322, Jun. 2010, doi: 10.1002/jmri.22004.
[14]A. Cerasaet al., “A Cellular Neural Network methodology for the automated segmentation of multiple sclerosis lesions,” J. Neurosci. Methods, vol. 203, no. 1, pp. 193–199, Jan. 2012, doi: 10.1016/j.jneumeth.2011.08.047.
[15]R. Simõeset al., “Automatic segmentation of cerebral white matter hyperintensities using only 3D FLAIR images,” Magn. Reson. Imaging, vol. 31, no. 7, pp. 1182–1189, Sep. 2013, doi: 10.1016/j.mri.2012.12.004.
[16]B. Xue, L. Wang, H.-C. Li, H. M. Chen, and C.-I. Chang, “Lesion detection in magnetic resonance brain images by hyperspectral imaging algorithms,” in Remotely Sensed Data Compression, Communications, and Processing XII, SPIE, Aug. 2016, pp. 106–117. doi: 10.1117/12.2223886.
[17]Malo Gaubert, Andrea Dell, Catharina Lange, Antoine Garnier-Crussard, Isabella Zimmermann, Martin Dyrba, Marco Duering, Gabriel Ziegler, Oliver Peters, Lukas Preis, Josef Priller, Eike Jakob Spruth, Anja Schneider,“Performance evaluation of automated WMHs seg-algor in a multicenter cohort on cognitive impairment and dementia_2023,”TYPE Original Research,2023, DOI 10.3389/fpsyt.2022.1010273.
[18]Soojin Lee, ZunHyanRieu, Regina EY Kim, Minho Lee,Kevin Yen, Junghyun Yong, DonghyeonKim“Automatic segmentation of WMHs in T2-FLAIR with AQUA_2023,”Brain Research Bulletin 205 (2023) 110825.
[19]J.-W. Chai et al., “Robust volume assessment of brain tissues for 3-dimensional fourier transformation MRI via a novel multispectral technique,” PloS One, vol. 10, no. 2, p. e0115527, 2015, doi: 10.1371/journal.pone.0115527.
[20]J.-W. Chai et al., “Quantitative analysis in clinical applications of brain MRI using independent component analysis coupled with support vector machine,” J. Magn. Reson. Imaging JMRI, vol. 32, no. 1, pp. 24–34, Jul. 2010, doi: 10.1002/jmri.22210.
[21]A. H. Kashani, M. Wong, N. Koulisis, C.-I. Chang, G. Martin, and M. S. Humayun, “Hyperspectral imaging of retinal microvascular anatomy,” J. Biomed. Eng. Inform., vol. 2, no. 1, p. 139, Nov. 2015, doi: 10.5430/jbei.v2n1p139.
[22]Y.-C. Ouyang et al., “Independent Component Analysis for Magnetic Resonance Image Analysis,” EURASIP J. Adv. Signal Process., vol. 2008, no. 1, pp. 1–14, Dec. 2008, doi: 10.1155/2008/780656.
[23]L. P. Clarke et al., “MRI segmentation: Methods and applications,” Magn. Reson. Imaging, vol. 13, no. 3, pp. 343–368, Jan. 1995, doi: 10.1016/0730-725X(94)00124-L.
[24]M. Vaidyanathan, L. P. Clarke, C. Heidtman, R. P. Velthuizen, and L. O. Hall, “Normal brain volume measurements using multispectral MRI segmentation,” Magn. Reson. Imaging, vol. 15, no. 1, pp. 87–97, 1997, doi: 10.1016/s0730-725x(96)00244-5.
[25]A. H. Andersen, Z. Zhang, M. J. Avison, and D. M. Gash, “Automated segmentation of multispectral brain MR images,” J. Neurosci. Methods, vol. 122, no. 1, pp. 13–23, Dec. 2002, doi: 10.1016/s0165-0270(02)00273-x.
[26]J. C. Harsanyi, “Detection and classification of subpixel spectral signatures in hyperspectral image sequences,” Ph.D. dissertation, Department of Electrical Engineering, University of Maryland, Baltimore County, United States, 1993.
[27]J. C. Harsanyi and C.-I. Chang, “Hyperspectral image classification and dimensionality reduction: an orthogonal subspace projection approach,” IEEE Trans. Geosci. Remote Sens., vol. 32, no. 4, pp. 779–785, Jul. 1994, doi: 10.1109/36.298007.
[28]H. Ren and C.-I. Chang, “A generalized orthogonal subspace projection approach to unsupervised multispectral image classification,” IEEE Trans. Geosci. Remote Sens., vol. 38, no. 6, pp. 2515–2528, Jan. 2000, doi: 10.1109/36.885199.
[29]C.-I. Chang et al., “Generalized constrained energy minimization approach to subpixel target detection for multispectral imagery,” Opt. Eng., vol. 39, no. 5, pp. 1275–1281, May 2000, doi: 10.1117/1.602486.
[30]Y.-C. Ouyang et al., “Band expansion-based over-complete independent component analysis for multispectral processing of magnetic resonance images,” IEEE Trans. Biomed. Eng., vol. 55, no. 6, pp. 1666–1677, Jun. 2008, doi: 10.1109/tbme.2008.919107.
[31]C.-M. Wang et al., “Unsupervised orthogonal subspace projection approach to magnetic resonance image classification,” Opt. Eng., vol. 41, no. 7, pp. 1546–1557, Jul. 2002, doi: 10.1117/1.1479710.
[32]C.-M. Wang et al., “Detection of spectral signatures in multispectral MR images for classification,” IEEE Trans. Med. Imaging, vol. 22, no. 1, pp. 50–61, Jan. 2003, doi: 10.1109/TMI.2002.806858.
[33]C. C.-C. Chen et al., “Intra-pixel multispectral processing of magnetic resonance brain images for tissue characterisation,” Int. J. Comput. Sci. Eng., vol. 8, no. 2, pp. 87–110, Jan. 2013, doi: 10.1504/IJCSE.2013.053090.
[34]G.-C. Lin, C.-M. Wang, W.-J. Wang, and S.-Y. Sun, “Automated classification of multispectral MR images using unsupervised constrained energy minimization based on fuzzy logic,” Magn. Reson. Imaging, vol. 28, no. 5, pp. 721–738, Jun. 2010, doi: 10.1016/j.mri.2010.03.009.
[35]S.-W. Chan et al., “Breast Tumor Detection and Classification Using Intravoxel Incoherent Motion Hyperspectral Imaging Techniques,” BioMed Res. Int., vol. 2019, p. e3843295, Jul. 2019, doi: 10.1155/2019/3843295.
[36]H.-M. Chen et al., “A Hyperspectral Imaging Approach to White Matter Hyperintensities Detection in Brain Magnetic Resonance Images,” Remote Sens., vol. 9, no. 11, p. 1174, Nov. 2017, doi: 10.3390/rs9111174.
[37]C.-C. C. Chen et al., “An Iterative Mixed Pixel Classification for Brain Tissues and White Matter Hyperintensity in Magnetic Resonance Imaging,” IEEE Access, vol. 7, pp. 124674–124687, 2019, doi: 10.1109/ACCESS.2019.2931761.
[38]C.-I. Chang, C.-C. C. CHEN, J. W. CHAI, and H.-M. CHEN, “Iterative analyzing method for a medical image,” US10621722B2, Apr. 14, 2020 Accessed: Mar. 13, 2024. [Online]. Available: https://patents.google.com/patent/US10621722B2/en
[39]X. Jiao and C.-I. Chang, “Kernel-based constrained energy minimization (K-CEM),” in Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIV, SPIE, Apr. 2008, pp. 523–533. doi: 10.1117/12.782221.
[40]K. Y. Ma and C.-I. Chang, “Kernel-Based Constrained Energy Minimization for Hyperspectral Mixed Pixel Classification,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–23, 2022, doi: 10.1109/TGRS.2021.3085801.
[41]C.-I. Chang, “An Effective Evaluation Tool for Hyperspectral Target Detection: 3D Receiver Operating Characteristic Curve Analysis,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 6, pp. 5131–5153, Jun. 2021, doi: 10.1109/TGRS.2020.3021671.
[42]C.-I. Chang, Hyperspectral Data Processing: Algorithm Design and Analysis. New York: John Wiley & Sons, Inc, 2013.
[43] O.L. Frost III, "An algorithm for linearly constrained adaptive array processing," Proc. IEEE, vol. 60, pp. 926-935, 1972.
[44] C. K. I. Williams and M. Seeger, “Using the Nyström method to speed up kernel machines,” in Proc. Neural Inf. Process. Syst. (NIPS), 2001, pp. 682–688.
[45] B. Xue, C. Yu, Y. Wang, M. Song, S. Li, L. Wang, H.M. Chen and C.-I Chang, “A subpixel target approach to hyperpsectral image classification,” IEEE Trans. on Geoscience and Remote Sensing, vol. 55, no. 9, pp. 5093-5114, September 2017.
[46]C.-I Chang, "Target signature-constrained mixed pixel classification for hyperspectral imagery," IEEE Trans. on Geoscience and Remote Sensing, vol. 40, no. 2, pp. 1065-1081, May 2002.
[47] H. Ren and C.-I Chang, "Target-constrained interference-minimized approach to subpixel target detection for hyperspectral imagery," Optical Engineering, vol. 39, no. 12, pp. 3138-3145, December 2000.
[48]Paul Schmidt. Bayesian inference for structured additive regression models for
large-scale problems with applications to medical imaging. PhD thesis, Ludwig- Maximilians-Universität München, Januar 2017. URL http://nbn-resolving.de/urn:nbn:de:bvb:19-203731.
[49]Paul Schmidt, Christian Gaser, Milan Arsic, Dorothea Buck, Annette Förschler,
Achim Berthele, Muna Hoshi, RüdigerIlg, Volker J Schmid, Claus Zimmer, et al.
An automated tool for detection of FLAIR-hyperintense white-matter lesions in
multiple sclerosis. Neuroimage, 59(4):3774–3783, 2012.
[50]Paul Schmidt, Viola Pongratz, Pascal Küster, Dominik Meier, Jens Wuerfel, Carsten Lukas, Barbara Bellenberg, Frauke Zipp, Sergiu Groppa, Philipp G Sämann, et al. Automated segmentation of changes in flair-hyperintense white matter lesions inmultiple sclerosis on serial magnetic resonance imaging. NeuroImage: Clinical,23:101849, 2019.