CHOSUN

클러스터링 기술과 워터쉐드 변환을 이용한 MRI 영상의 효율적인 세션화

Metadata Downloads
Author(s)
바수카라 디바쉬
Issued Date
2016
Keyword
Image segmentation, Watershed transformation, Clustering, Thresholding, Markers
Abstract
Watershed transformation is an effective segmentation algorithm that originates from the mathematical morphology. This algorithm is widely used in medical image segmentation because it produces complete division even under poor contrast. However, over-segmentation is its most significant limitation. Therefore, this thesis proposes a combination of watershed transformation and the clustering algorithm to segment magnetic resonance brain images efficiently. The clustering algorithm is used to form clusters. Then, the brightest cluster which contains gray matter (GM) and cerebrospinal fluid (CSF) is thus selected and converted into a binary image. A Sobel operator applied on the binary image generates the initial gradient magnitude image. Morphological image reconstruction is applied to find the foreground and background markers. The final gradient magnitude image is obtained using the minima imposition technique on the initial gradient magnitude along with markers. In addition, watershed segmentation applied on the final gradient magnitude generates effective GM and CSF segmentation. The results are compared with simple marker-controlled watershed segmentation, watershed segmentation combined with Otsu multilevel thresholding, and distance regularized level set for validation.|워터쉐드 변환은 수학적 형태학에서 유래한 효과적인 알고리즘이다. 이 알고리즘은 낮은 명암에서도 효과적으로 분할되어 의료 영상 분할에 폭넓게 활용된다. 하지만 의료 영상에서의 과분할은 가장 중요한 제한 사항이다. 제안하는 방법은 워터쉐드 변환과 클러스터링 알고리즘의 조합을 통한 자기 공명 뇌 영상의 효율적인 분할이다. 우선 클러스터링 알고리즘은 특징 집합을 형성하는데 사용된다. 이후, 높은 명암 부분의 클러스터에 포함되는 회백질(Gray-Matter, GM)과 뇌척수액(Cerebro Spinal Fluid, CSF) 영역을 선택하여 이진 영상으로 변환한다. 변환 된 이진 영상은 Sobel 연산을 통해 초기 기울기 크기 영상을 생성하고 형태적(Morphological)영상 복원은 전경과 배경 마커를 찾기 위해 사용한다. 최종 기울기 크기 영상은 마커와 함께 얻어지며, 이는 초기 기울기 크기값에 최소값 부과 기법을 적용하여 얻을 수 있다. 또한 워터쉐드 분할에 최종 기울기 크기를 적용하여 회백질과 뇌척수액을 효과적으로 분할한다. 제안한 방법의 실험 결과는 심플 마커 제어 워터쉐드 분할, Otsu Multilevel 임계값과 결합된 워터쉐드 변환 그리고 거리 정규화 레벨 셋과 비교 분석하였다.
Alternative Title
An Efficient Segmentation of Magnetic Resonance Brain Image using Clustering Technique and Watershed Transform
Alternative Author(s)
Dibash Basukala
Affiliation
정보통신공학과
Department
일반대학원 정보통신공학과
Advisor
권구락
Awarded Date
2016-08
Table Of Contents
List of Figures iii
List of Tables v
Acronyms vi
Abstract vii
Abstract Korean viii

1 INTRODUCTION 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Problems and Challenges of Brain Image Segmentation . . . . . 3
1.3 Scopes and Objectives of the Thesis . . . . . . . . . . . . . . . 3
1.4 Thesis Contribution . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 IMAGING MODALITIES 5
2.1 Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . 5
3 SEGMENTATION METHODS 7
3.1 Manual Segmentation . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Region Growing . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.4 Clustering Algorithms . . . . . . . . . . . . . . . . . . . . . . 10
3.4.1 K-means Clustering Algorithm . . . . . . . . . . . . . 11
3.4.2 Fuzzy C-Means Clustering . . . . . . . . . . . . . . . . 12
3.5 Watershed Transform . . . . . . . . . . . . . . . . . . . . . . . 13
3.5.1 Rainfall Approach . . . . . . . . . . . . . . . . . . . . 14
3.5.2 Flooding Approach . . . . . . . . . . . . . . . . . . . . 14
3.6 Atlas-Guided Approaches . . . . . . . . . . . . . . . . . . . . . 16
3.7 Active Contours . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.8 Multiphase Active Contours . . . . . . . . . . . . . . . . . . . 18
3.9 Hybrid Segmentation Methods . . . . . . . . . . . . . . . . . . 18
4 THE PROPOSED METHOD 20
4.1 Filtering Operation . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2 Expectation-Maximization Algorithm . . . . . . . . . . . . . . 21
4.3 Thresholding . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.4 Edge Detection Method . . . . . . . . . . . . . . . . . . . . . . 25
4.5 Morphological Image Reconstruction . . . . . . . . . . . . . . 27
4.5.1 Dilation . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.5.2 Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.5.3 Erosion Based Gray-Scale Image Reconstruction . . . . 28
4.5.4 Dilation Based Gray-Scale Image Reconstruction . . . . 28
4.6 Markers Extraction . . . . . . . . . . . . . . . . . . . . . . . . 29
5 PERFORMANCE EVALUATION 30
5.1 Subjective Quality . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2 Segmentation Validation and Quantitative Analysis . . . . . . . 39
5.2.1 Success Rates . . . . . . . . . . . . . . . . . . . . . . . 40
5.2.2 Similarity Metrics . . . . . . . . . . . . . . . . . . . . 41
6 CONCLUSIONS 46
Degree
Master
Publisher
조선대학교
Citation
바수카라 디바쉬. (2016). 클러스터링 기술과 워터쉐드 변환을 이용한 MRI 영상의 효율적인 세션화.
Type
Dissertation
URI
https://oak.chosun.ac.kr/handle/2020.oak/12972
http://chosun.dcollection.net/common/orgView/200000265814
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.