CHOSUN

뇌 MRI 와 SVM 분류기로부터 낮은 차원의 특징을 이용한 알츠하이머 병 예측 기술

Metadata Downloads
Author(s)
알람 사루아르
Issued Date
2016
Keyword
AD, CWT, DWT, DtCWT, RBF-SVM, MRMR, PCA
Abstract
Alzheimer’s disease (AD) is the most frequent form of dementia, causes more health and socioeconomic problem. This progressive neurodegenerative disorder causes damage of brain cells, affects cognitive acts, behavioral problems, and memory disorder. At present, a lot of computer-aided diagnosis (CAD) based research is going on using Magnetic Resonance imaging (MRI) as a biomarker. Early diagnosis could help the patient to take preventive cure to get rid of AD risk factors to generate further. Around 50% MCI (Mild Cognitive Impairment) patient develop AD in three to four years. In this paper, a novel method is proposed to predict AD from normal controls (NC) here. In this work, maximum relevant minimum redundant principal components of dual tree complex wavelet transform (dtCWT) coefficients are extracted. The transaxial slices of MR images are selected for extracting dtCWT coefficients here. After linear discriminant analysis (LDA) of those coefficients, kernel SVM is trained and tested. The accuracy, sensitivity, and specificity we have achieved using proposed approach are comparable or superior to those obtained by various conventional AD prediction methods found in the literature.|알츠하이머성 치매는(AD) 치매의 가장 흔한 형태이며, 치료 비용으로 인한 경제적 문제와 더불어 사화적 문제가 발생한다. 이는 뇌 세포 손상으로 인한 신경 퇴행성 장애로서 인지장애, 행동장애, 기억장애를 유발한다. 현재 컴퓨터를 이용한 진단으로써 많은 연구가 이루어지며 대표적으로 자기 공명 영상(MRI)의 바이오마커를 이용하여 진단하는 방법이 있다. 치매 조기 진단은 알츠하이머성 치매의 초기치료를 통해 추가적으로 환자가 늘어나지 않게 조치 할 수 있다. 50%의 MCI(경도인지장애) 환자는 3~4 년 동안 알츠하이머성 치매가 진행된다. 본 논문에서는 정상인 대조군(NC)에서 알츠하이머성 치매를 예측하는 방법을 제안한다. 본 연구에서는, dtCWT(듀얼 트리 콤플렉스 웨이블릿 변환)의 최대에서 최소한의 중복 주성분 계수를 추출한다. 뇌 MRI 영상의 시상면, 관상면, 가로면(Transverse Plane) 영상은 dtCWT 계수를 추출하기 위해 사용된다. 이후 계수들은 LDA(선형 판별 분석)를 거쳐 Kernal-SVM 학습을 통해 트레이닝되고 테스트된다. 이 논문에서 제안하는 방법은 기존의 다양한 연구들에서 찾을 수 있는 AD 예측방법과 정확성, 민감도, 특이도를 따졌을 때 더욱 우수하거나 대등한 성능을 보인다.
Alternative Title
Alzheimer Disease Classification Using Lower Dimensional Features from Brain MRI and SVM Classifiers
Alternative Author(s)
Saruar Alam
Department
일반대학원 정보통신공학과
Advisor
권구락
Awarded Date
2016-08
Table Of Contents
Table of Contents i
List of Tables iii
List of Figures iv
Acronyms vi
Abstract vii
1 Introduction 1
1.1 Thesis motivation and overview . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Goals of this work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Background 5
2.1 Alzheimer disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Magnetic Resonance Imaging Image . . . . . . . . . . . . . . . . . . 8
2.3 MR Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1 FreeSurfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.2 Discrete Wavelet Transform . . . . . . . . . . . . . . . . . . 16
2.3.3 Feature selection . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3.4 Dimensionality reduction technique . . . . . . . . . . . . 21
2.4 Support Vector Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3. The proposed method 24
3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Dual Tree Complex Wavelet Transform . . . . . . . . . . . . . . . 24
3.3 Max-Relevance and Min-Redundancy feature subset . . . . 28
3.4 Principal component Analysis . . . . . . . . . . . . . . . . . . . . . . 29
3.5 Linear discriminant Analysis . . . . . . . . . . . . . . . . . . . . . . . 31
3.6 Kernel Support Vector Machine . . . . . . . . . . . . . . . . . . . . . 32
4. Performance Evaluation 35
4.1 Accuracy, Sensitivity, Specificity . . . . . . . . . . . . . . . . . . . . . . . 35
4.2 Cross validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5. Performance Analysis 40
5.1 Overview of experimental data . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.2 Dual Tree Complex Wavelet Transform feature . . . . . . . . . . . 41
5.3 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6. Conclusion 52
Degree
Master
Publisher
조선대학교
Citation
알람 사루아르. (2016). 뇌 MRI 와 SVM 분류기로부터 낮은 차원의 특징을 이용한 알츠하이머 병 예측 기술.
Type
Dissertation
URI
https://oak.chosun.ac.kr/handle/2020.oak/12887
http://chosun.dcollection.net/common/orgView/200000265658
Appears in Collections:
General Graduate School > 3. Theses(Master)
Authorize & License
  • AuthorizeOpen
  • Embargo2016-08-25
Files in This Item:

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.