發(fā)布時間：2018-07-07 14:36

  本文選題：功能磁共振成像 + 結(jié)構(gòu)磁共振成像��； 參考：《北京理工大學(xué)》2015年博士論文

【摘要】：近年來,MRI圖像分析方法被越來越多地應(yīng)用于大腦結(jié)構(gòu)和功能研究以及神經(jīng)性疾病的計(jì)算機(jī)輔助診斷中。另一方面,隨著人工智能的發(fā)展,機(jī)器學(xué)習(xí)技術(shù),尤其是其中的稀疏特征學(xué)習(xí)也被越來越多地引入MRI圖像分析中,在分類和預(yù)測建模方面扮演著重要角色。因此,研究新的MRI圖像分析方法,是更好地深度挖掘MRI圖像信息,進(jìn)而促進(jìn)腦科學(xué)研究及計(jì)算機(jī)輔助診斷技術(shù)發(fā)展的關(guān)鍵。然而,MRI圖像分析常常面臨小樣本、高特征維度的問題,由此導(dǎo)致的過擬合、噪聲特征和冗余特征嚴(yán)重降低了模型性能。稀疏特征學(xué)習(xí)方法能夠很好地解決上述難點(diǎn),并已經(jīng)成功地應(yīng)用于信號處理、模式識別和計(jì)算機(jī)視覺領(lǐng)域。本論文致力于研究新的適用于MRI圖像分析的稀疏特征學(xué)習(xí)方法,通過設(shè)計(jì)新的代價函數(shù)約束項(xiàng)來挑選擁有最佳分辨性能的圖像特征,從而提升模型分類和預(yù)測性能。本研究涉及的稀疏特征學(xué)習(xí)方法既包括稀疏單任務(wù)學(xué)習(xí),如稀疏貝葉斯學(xué)習(xí)和基于1L范數(shù)的稀疏學(xué)習(xí),也包括稀疏多任務(wù)學(xué)習(xí),如組Lasso,Dirty model和稀疏組Lasso。通過對上述方法進(jìn)行不同程度的創(chuàng)新,將其應(yīng)用于認(rèn)知神經(jīng)科學(xué)和神經(jīng)疾病診斷研究中,取得了良好的效果。本文的工作及創(chuàng)新之處主要包括以下5個部分:1.建立了一個基于多體素模式分析的學(xué)習(xí)模型,在初級視皮層上對空間視覺刺激進(jìn)行解碼研究,解決了單體素分析方法忽視了體素之間的相關(guān)信息這一缺點(diǎn)。進(jìn)一步,建立了一種多分類的稀疏貝葉斯學(xué)習(xí)模型,將特征選擇與視覺解碼結(jié)合起來。該模型能夠在選擇最相關(guān)特征的同時利用挑選出的特征進(jìn)行視覺解碼,具有很好的整合性。實(shí)驗(yàn)結(jié)果表明,該方法從2000個初級視皮層體素中挑選出9個最相關(guān)體素,使用挑選出的體素進(jìn)行解碼,分類精度達(dá)到91.6%。同時,將挑選出的9個體素映射回原始腦空間,從另一個角度驗(yàn)證了初級視皮層具有視網(wǎng)膜映射特性。該方法為基于功能MRI的視覺研究提供了一種新的途徑。2.首次提出一種基于結(jié)構(gòu)MRI圖像和1L范數(shù)稀疏特征學(xué)習(xí)的煙霧病診斷方法,解決了傳統(tǒng)數(shù)字造影方法有損、技術(shù)復(fù)雜及代價高昂的缺點(diǎn),使得將煙霧病診斷作為常規(guī)檢查成為可能。具體來說,該方法首先提取結(jié)構(gòu)MRI圖像的皮層厚度特征,每幅圖像得到約2萬個特征。然后建立了三種基于1L范數(shù)的稀疏特征學(xué)習(xí)模型,包括Lasso,彈性網(wǎng)和L1-logistic回歸,通過特征選擇實(shí)現(xiàn)特征約簡,最后用挑選出的特征訓(xùn)練支持向量機(jī)分類器。實(shí)驗(yàn)結(jié)果表明,提出的診斷方法取得了較好的診斷精度(分類精度),其中基于彈性網(wǎng)特征學(xué)習(xí)的方法取得了最高的診斷精度,達(dá)到82.36%,對應(yīng)ROC曲線下的面積0.833,顯著優(yōu)于未經(jīng)過特征選擇直接使用支持向量機(jī)對所有提取特征分類的結(jié)果(分類精度71.72%,對應(yīng)ROC曲線下面積0.787)。3.利用兒童(6歲到15歲)的結(jié)構(gòu)MRI圖像,建立了基于多核支持向量回歸的智商估計(jì)模型,并在建模過程中提出一種改進(jìn)Dirty model多任務(wù)特征學(xué)習(xí)方法用于特征選擇,較好地實(shí)現(xiàn)了對兒童的智商估計(jì)。具體來說,首先提取兒童結(jié)構(gòu)MRI圖像的灰質(zhì)/白質(zhì)特征,將對灰質(zhì)/白質(zhì)的特征選擇分別看做一個學(xué)習(xí)任務(wù),利用提出的改進(jìn)Dirty model選擇與智商相關(guān)的灰質(zhì)/白質(zhì)特征。分別計(jì)算挑選出的灰質(zhì)/白質(zhì)特征的核函數(shù),送入多核支持向量回歸模型中進(jìn)行智商估計(jì)。實(shí)驗(yàn)結(jié)果顯示,提出的方法估計(jì)的智商分?jǐn)?shù)與兒童真實(shí)智商分?jǐn)?shù)的相關(guān)性為0.718,與真實(shí)智商分?jǐn)?shù)之間的均方根誤差為8.695。該智商估計(jì)模型的意義在于,為今后預(yù)測嬰幼兒智商,從而根據(jù)其預(yù)測智商定制合理的早期學(xué)習(xí)計(jì)劃提供了方法學(xué)基礎(chǔ)。4.提出一種基于分層模型和改進(jìn)Dirty model的多被試解碼方法,較好地解決了傳統(tǒng)基于功能MRI解碼方法需要對每個被試單獨(dú)建模的缺點(diǎn)。傳統(tǒng)的解碼方法以體素作為特征,但不同被試之間對于相同實(shí)驗(yàn)刺激的體素激活模式差異較大,難以對所有被試建立一個統(tǒng)一的模型。為解決這一問題,提出一種分層模型,以體素作為低層特征,學(xué)習(xí)出更加魯棒的高層特征,然后利用提出的改進(jìn)Dirty model對高層特征進(jìn)行選擇,并使用選擇出的高層特征,對所有被試建立一個統(tǒng)一的解碼模型。將提出的多被試解碼方法用于2D/3D視覺刺激的功能MRI數(shù)據(jù)進(jìn)行驗(yàn)證,分類精度達(dá)到89.4%,顯著高于直接使用體素作為特征的方法(73.4%)。該方法為研究不同被試間的神經(jīng)特性提供了方法學(xué)上的支持。5.提出一種基于結(jié)構(gòu)MRI圖像和稀疏特征學(xué)習(xí)的自閉癥診斷方法,為解決傳統(tǒng)自閉癥診斷方法以行為學(xué)評分為標(biāo)準(zhǔn),而實(shí)際中很難用一種行為學(xué)評分去診斷該疾病這一問題提供了一種新的方法學(xué)途徑,并在建模過程中提出一種全新的Canonical圖匹配稀疏組Lasso多任務(wù)特征學(xué)習(xí)算法用于特征選擇。具體地說,首先提取結(jié)構(gòu)MRI圖像的灰質(zhì)/白質(zhì)特征,通過典型相關(guān)分析將原始灰質(zhì)/白質(zhì)特征映射到一個新的Canonical空間。將對類標(biāo)簽以及SRS_TOTAL行為學(xué)評分的Canonical特征選擇分別看做一個學(xué)習(xí)任務(wù),利用提出的方法選擇與自閉癥分類最相關(guān)的Canonical特征,然后用支持向量機(jī)進(jìn)行分類。實(shí)驗(yàn)結(jié)果顯示,提出的方法診斷精度為75.4%,ROC曲線下面積為0.804,顯著優(yōu)于基于原始灰質(zhì)/白質(zhì)特征的最新方法,也優(yōu)于現(xiàn)有的其它基于Canonical特征的方法。
[Abstract]:In recent years, MRI image analysis methods have been used more and more in the study of brain structure and function and computer aided diagnosis of neuropathic diseases. On the other hand, with the development of artificial intelligence, machine learning technology, especially the sparse feature learning among them, has also been introduced into MRI image analysis more and more, in classification and prediction Modeling plays an important role. Therefore, the study of the new MRI image analysis method is the key to the better depth mining of MRI image information, thus promoting the development of brain science research and computer aided diagnosis technology. However, MRI image analysis often faces small samples, high characteristic dimensions, resulting in overfitting, noise characteristics and redundancy. The redundant features seriously reduce the performance of the model. The sparse feature learning method can solve the above difficulties well, and has been successfully applied to the field of signal processing, pattern recognition and computer vision. This paper is devoted to the study of the new sparse feature learning method for MRI image analysis and the design of new cost function constraints. The sparse feature learning methods involved in this study include sparse single task learning, such as sparse Bayesian learning and sparse learning based on 1L norm, and sparse multitask learning, such as group Lasso, Dirty model, and sparse group Lasso. through pairs. This method has been applied to cognitive neuroscience and neural disease diagnosis research and has achieved good results. The work and innovation of this paper mainly include the following 5 parts: 1. a learning model based on the model analysis of the multibody element is established to decode the spatial visual stimuli on the primary visual cortex. In addition, a multi classification sparse Bayesian learning model is established, which combines feature selection with visual decoding. The model can be used to select the most relevant features and use the selected features for visual decoding, which is very good. The experimental results show that the method selects 9 most relevant voxels from 2000 primary visual cortex voxels, decodes the selected voxels, and the classification accuracy reaches 91.6%.. The selected 9 individual elements are mapped back to the original brain space and the retina mapping characteristics of the primary visual cortex are verified from another angle. The method provides a new way for visual research based on functional MRI..2. first proposes a method of diagnosis of moyamoya disease based on structural MRI image and 1L norm sparse feature learning. It solves the disadvantages of traditional digital contrast method, which is complicated and costly. It makes it possible to make the diagnosis of smoke fog as a routine examination. For example, this method first extracts the cortical thickness characteristics of the structure MRI image, and each image gets about 20 thousand features. Then three sparse feature learning models based on 1L norm are established, including Lasso, elastic network and L1-logistic regression. Feature reduction is implemented by feature selection. Finally, the selected feature is used to train the support vector machine classifier. The experimental results show that the proposed method has achieved good diagnostic accuracy (classification accuracy), and the method based on the feature learning of the elastic network has achieved the highest diagnostic accuracy, reached 82.36%, corresponding to the area of 0.833 under the ROC curve, and is significantly better than the node without the feature selection using the support vector machine to classify all the extracted features. (classification accuracy 71.72%, corresponding to the area under the ROC curve 0.787).3. using the structure MRI images of children (6 to 15 years old), an IQ estimation model based on multi kernel support vector regression is established. In the modeling process, an improved Dirty model multi task feature learning method is proposed for feature selection, and the IQ estimation of children is better realized. Specifically, the gray matter / white matter features of the children's structure MRI images are first extracted and the selection of gray matter / white matter features is considered as a learning task respectively. The gray matter / white matter characteristics associated with the IQ are selected by the proposed improved Dirty model. The kernel functions of the selected gray matter / white matter are calculated and sent to the multicore support vector back. The results show that the correlation between the estimated IQ score and the true IQ score of the proposed method is 0.718, and the root mean square error between the true IQ score and the true IQ score is 8.695., the IQ estimation model is for the future prediction of infant IQ, so that it is reasonable to customize the IQ according to its IQ prediction. The early learning plan provides a methodological basis for.4. to propose a multi decode method based on a hierarchical model and an improved Dirty model, which is a good solution to the shortcomings of the traditional functional MRI decoding method that needs to be modeled individually for each subject. The traditional decoding method is characterized by voxels, but different subjects have the same experimental stings between the different subjects. In order to solve this problem, a hierarchical model is proposed to learn more robust high level features, and then use the proposed improved Dirty model to select the high level features and use the selected high-level features. A unified decoding model is established for all subjects. The proposed multi test decoding method is used to verify the functional MRI data of 2D/3D visual stimuli. The classification accuracy is 89.4%, which is significantly higher than that of the direct use of voxel (73.4%). This method provides a methodological support for the study of the neural characteristics between different subjects. An autism diagnosis method based on structural MRI image and sparse feature learning is proposed to solve the traditional autism diagnosis method, which is based on behavioral score. In practice, it is difficult to use a kind of behavioral score to diagnose the disease. A new way is provided in the process of modeling, and a new Cano is proposed in the modeling process. Nical graph matching sparse group Lasso multi task feature learning algorithm is used for feature selection. Specifically, first extract gray / white texture features of structural MRI images, map the original gray / white matter features to a new Canonical space through canonical correlation analysis, and select the Canonical features of class labels and SRS_TOTAL behavioral scores. Do not look at a learning task, use the proposed method to select the most relevant Canonical features with the autism classification, and then use the support vector machine to classify them. The experimental results show that the proposed method has a diagnostic accuracy of 75.4% and the area under the ROC curve is 0.804, which is significantly better than the latest method based on the original gray matter / white matter characteristics, and is better than the existing one. Other methods based on Canonical features.
【學(xué)位授予單位】：北京理工大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2015
【分類號】：TP391.41

【共引文獻(xiàn)】

相關(guān)期刊論文 前10條

1 唐曉芬;趙秉新;;基于支持向量機(jī)的農(nóng)村勞動力轉(zhuǎn)移預(yù)測[J];安徽農(nóng)業(yè)科學(xué);2011年11期

2 郭良棟;田江;叢曉東;;基于CPSO和SVM的混沌時間序列預(yù)測[J];遼寧科技大學(xué)學(xué)報;2009年06期

3 李冬;王璐;康文崢;;基于SVM的作戰(zhàn)效能靈敏度分析[J];兵工自動化;2011年01期

4 張軍;單永海;曹殿廣;鄭玉新;葛欣鑫;;基于最小二乘支持向量機(jī)的機(jī)槍加速壽命建模[J];兵工學(xué)報;2012年01期

5 龐九鳳;李險峰;謝勁松;佟冬;程旭;;基于支持向量機(jī)的微體系結(jié)構(gòu)設(shè)計(jì)空間探索(英文)[J];北京大學(xué)學(xué)報(自然科學(xué)版);2010年01期

6 廖龍飛;王晶;;面向多輸出系統(tǒng)的啟發(fā)式支持向量機(jī)回歸[J];北京化工大學(xué)學(xué)報(自然科學(xué)版);2011年02期

7 ;Application of support vector machine in the prediction of mechanical property of steel materials[J];Journal of University of Science and Technology Beijing(English Edition);2006年06期

8 王澤生;車喜龍;;無線接入點(diǎn)吞吐率預(yù)測的建模與優(yōu)化[J];吉林大學(xué)學(xué)報(信息科學(xué)版);2010年03期

9 ;Nonlinear Time Series Prediction Using LS-SVM with Chaotic Mutation Evolutionary Programming for Parameter Optimization[J];Communications in Theoretical Physics;2006年04期

10 ;Online Dispatching Decision Support System for Smart Grid Reservoir Group[J];Electricity;2011年05期

相關(guān)會議論文 前10條

1 劉卓軍;李曉明;;一種基于時間序列異常點(diǎn)監(jiān)測的可疑交易分析方法[A];第十三屆中國管理科學(xué)學(xué)術(shù)年會論文集[C];2011年

2 顏七笙;;基于PCA-SVM的動態(tài)聯(lián)盟候選伙伴績效評價方法[A];第十三屆中國管理科學(xué)學(xué)術(shù)年會論文集[C];2011年

3 趙衛(wèi);劉濟(jì)科;;隨機(jī)結(jié)構(gòu)可靠度分析的支持向量機(jī)方法[A];第九屆全國振動理論及應(yīng)用學(xué)術(shù)會議論文集[C];2007年

4 周金柱;黃進(jìn);甄立冬;;腔體濾波器制造精度對電性能影響的支持向量建模與優(yōu)化[A];2011年機(jī)械電子學(xué)學(xué)術(shù)會議論文集[C];2011年

5 ;Online Modeling Based on Support Vector Machine[A];2009中國控制與決策會議論文集（1）[C];2009年

6 ;An online outlier detection method for process control time series[A];Proceedings of the 2011 Chinese Control and Decision Conference（CCDC）[C];2011年

7 劉子陽;郭崇慧;;應(yīng)用支持向量回歸方法預(yù)測胎兒體重[A];大連理工大學(xué)生物醫(yī)學(xué)工程學(xué)術(shù)論文集（第2卷）[C];2005年

8 徐慧;李云瑋;;基于分形和支持向量機(jī)的瓦斯涌出量預(yù)測研究[A];煤礦自動化與信息化——第19屆全國煤礦自動化與信息化學(xué)術(shù)會議暨中國礦業(yè)大學(xué)（北京）百年校慶學(xué)術(shù)會議論文集[C];2009年

9 印家健;李夢龍;;基于HMLP參數(shù)的二肽logP值支持向量機(jī)預(yù)測模型[A];第九屆全國計(jì)算（機(jī)）化學(xué)學(xué)術(shù)會議論文摘要集[C];2007年

10 于國榮;夏自強(qiáng);;支持向量機(jī)的混沌序列預(yù)測模型及在徑流中應(yīng)用[A];2007重大水利水電科技前沿院士論壇暨首屆中國水利博士論壇論文集[C];2007年

相關(guān)博士學(xué)位論文 前10條

1 王曉明;基于統(tǒng)計(jì)學(xué)習(xí)的模式識別幾個問題及其應(yīng)用研究[D];江南大學(xué);2010年

2 吳少智;時間序列數(shù)據(jù)挖掘在生物醫(yī)學(xué)中的應(yīng)用研究[D];電子科技大學(xué);2010年

3 祝曉春;增強(qiáng)型軟件項(xiàng)目中測試工作量度量研究[D];浙江大學(xué);2010年

4 劉春波;統(tǒng)計(jì)建模方法的理論研究及應(yīng)用[D];江南大學(xué);2011年

5 劉雪美;噴桿噴霧機(jī)風(fēng)助風(fēng)筒多目標(biāo)優(yōu)化設(shè)計(jì)[D];山東農(nóng)業(yè)大學(xué);2010年

6 王中鋒;樹型貝葉斯網(wǎng)絡(luò)分類器鑒別式訓(xùn)練研究[D];北京交通大學(xué);2011年

7 鄧鵬程;基于數(shù)據(jù)的鉛鋅熔煉過程自適應(yīng)在線監(jiān)控與故障診斷[D];中南大學(xué);2011年

8 李小敏;胎肺超聲圖像分析及其在肺成熟度估計(jì)的應(yīng)用[D];復(fù)旦大學(xué);2011年

9 杜健輝;基于旋轉(zhuǎn)電弧傳感的水下焊縫成形及自動跟蹤的研究[D];華南理工大學(xué);2011年

10 左琳;神經(jīng)網(wǎng)絡(luò)及在網(wǎng)絡(luò)用戶行為分析中的應(yīng)用研究[D];電子科技大學(xué);2011年

相關(guān)碩士學(xué)位論文 前10條

1 姜成玉;基于支持向量機(jī)的時間序列預(yù)測[D];遼寧師范大學(xué);2010年

2 徐俏;基于多目標(biāo)微粒群的轉(zhuǎn)爐合金加入量優(yōu)化研究[D];大連理工大學(xué);2010年

3 姜力文;基于信息熵的轉(zhuǎn)爐煉鋼動態(tài)過程建模研究[D];大連理工大學(xué);2010年

4 陳啟;基于支持向量回歸在短期負(fù)荷預(yù)測中的應(yīng)用[D];河北工程大學(xué);2010年

5 陳逸波;支持向量機(jī)在高爐爐溫預(yù)報中的應(yīng)用及若干改進(jìn)[D];浙江大學(xué);2010年

6 王霏;面向動漫素材的特征提取與分類識別[D];浙江大學(xué);2011年

7 榮正勤;甜葉菊糖甙含量近紅外光譜定量預(yù)測模型的創(chuàng)建及應(yīng)用[D];浙江大學(xué);2011年

8 陳麗娟;利用對接分?jǐn)?shù)預(yù)測人類血漿白蛋白結(jié)合率[D];浙江大學(xué);2011年

9 李一磊;服裝設(shè)計(jì)風(fēng)格決策模型的研究與實(shí)現(xiàn)[D];東華大學(xué);2011年

10 韓國軍;基于特征選擇和支持向量機(jī)的紋理圖像分類[D];西安電子科技大學(xué);2009年

，

本文編號：2105263