흉부 CT 영상에서 밝기값 제약 기반 다중변형모델을 이용한 간유리음영 결절 자동 분할 = Automatic Segmentation of Ground-glass Opacity Nodule using Multi-phase Deformable Model with Intensity Constraint in Chest CT Images

국문 초록 (Abstract)

본 논문에서는 흉부 CT 영상에서 간유리음영 성분과 고형 성분을 배경과 잘 분리되도록 동시에 분할하는 방법을 제안한다. 제안하는 방법은 다음의 세 단계로 구성된다. 첫째, 밝기값 기반 히스토그램 모델링을 이용해 적응적 임계값을 산정하여 간유리음영 성분과 고형 성분의 초기 영역을 분할한다. 둘째, 유사한 밝기값을 갖는 주변기관이 함께 분할된 초기 영역을 보정하기 위하여 형태 정보와 폐분할 정보를 이용하여 결절 주위의 혈관과 흉벽을 제거한다. 셋째, 밝기값 제약 기반의 다중변형모델을 이용하여 간유 리음영 성분과 고형 성분 영역을 최종 분할한다. 제안방법의 성능을 평가하기 위해 기존 방법과 제안 방법의 분할 결과를 임상의에 의한 수동 분할 결과와 비교한다. 실험 결과는 제안 방법이 간유리음영 성분과 배경을 잘 구분하였음을 보여주며, 제안 방법의 정확성이 기존 방법 대비 77% 향상되었음을 보여준다.

다국어 초록 (Multilingual Abstract)

In this paper, we propose a simultaneous segmentation method to separate ground-glass opacity(GGO) and solid components from background in chest CT images. The proposed method consists of following three steps. First, initial GGO and solid components are segmented by estimating the adaptive threshold values using intensity histogram modeling. Second, to refine the initial segmentation that includes neighbor structures with similar intensities, vessels and the chest wall around the nodule are eliminated by using shape information and segmented lung information. Third, final GGO and solid components are segmented by using multi-phase deformable model with intensity constraint. To evaluate performance of the proposed method, we compare the segmentation results of conventional method and proposed method with manual segmentation result defined by radiologist. The results show that the proposed method accurately separates GGO component from background and the accuracy of the proposed method is improved by 77% than conventional method.

목차 (Table of Contents)

• 요약
• Abstract
• 1. 서론
• 2. 간유리음영 및 고형 성분 자동 분할
• 3. 실험 및 결과
• 4. 결론
• References

참고문헌 (Reference)

1 조현희, "폐 CT 영상에서 밝기값, 곡률 및 형태 정보를 이용한 결절 자동 분할" 한국정보과학회 39 (39): 537-546, 2012

2 B. L. Odry., "Solid Component Evaluation in Mixed Ground Glass Nodules" 6512 : 2007

3 Y. Yoo., "Segmentation of Ground Glass Opacities by Asymmetric Multi-phase Deformable Model" 6144 : 2006

4 P. Perona., "Scale-space and edge detection using anisotropic diffusion" 12 (12): 629-639, 1990

5 T. Kubota., "Robust Segmentation of Pulmonary Nodules of Various Densities : From Ground-glass Opacities to Solid Nodules" 253-262, 2008

6 K. Okada., "Robust 3D Segmentation of Pulmonary Nodules in Multislice CT Images" 881-889, 2004

7 J. M. Goo., "Ground-Glass Nodules on Chest CT as Imaging Biomarkers in the Management of Lung Adenocarninoma" 196 (196): 533-543, 2010

8 Y. Yim., "Correction of Segmented Lung Boundary for Inclusion of Pleural Nodules and Pulmonary Vessel in Chest CT Images" 38 (38): 845-857, 2008

9 H. Kobatake., "Convergence Index Filter for Vector Fields" 8 (8): 1029-1038, 1999

10 L. Zhang., "Consistent Interactive Segmentation of Pulmonary Ground Glass Nodules Identified in CT Studies" 5370 : 1709-1701, 2004

11 S. Oda., "Computer-Aided Volumetry of Plumonary Nodules Exhibiting Ground-Glass Opacity at MDCT" 194 (194): 398-406, 2010

12 R. Tachibana., "Automatic Segmentation of Pulmonary Nodules on CT Images by use of NCI Lung Image Database Consortium" 6144 : 2008

13 X. Ye., "Automatic Graph Cut Segmentation of Lesions in CT using Mean Shift Super-Pixels" 2010

14 W. A. Browder., "Automated Volumetric Segmentation Method for Growth Consistency of Nonsolid Pulmonary Nodules in High-resolution" 6514 : 2007

15 W. Son., "Automated Volumetric Segmentation Method for Computerized-Diagnosis of Pure Nodular Groundglass Opacity in High-resolution CT" 7624 : 2010

16 L. Zhang., "A Computer-based Method of Segmenting Ground Glass Nodules in Pulmonary CT Images : Comparison to Expert Radiologists’ Interpretations" 5747 : 113-123, 2005

17 S. D. Nie., "A CI Featurebased Pulmonary Nodule Segmentation using Three-domain Mean Shift Clustering" 223-227, 2007