Journal of the Korea Computer Graphics Society (한국컴퓨터그래픽스학회논문지)

Korea Computer Graphics Society (한국컴퓨터그래픽스학회)
권호 표지
DOI QR코드

DOI QR Code

Automated Classification of Ground-glass Nodules using GGN-Net based on Intensity, Texture, and Shape-Enhanced Images in Chest CT Images

흉부 CT 영상에서 결절의 밝기값, 재질 및 형상 증강 영상 기반의 GGN-Net을 이용한 간유리음영 결절 자동 분류

  • Byun, So Hyun (Department of Software Convergence, Seoul Women's University) ;
  • Jung, Julip (Department of Software Convergence, Seoul Women's University) ;
  • Hong, Helen (Department of Software Convergence, Seoul Women's University) ;
  • Song, Yong Sub (Department of Radiology, Seoul National University College of Medicine, and Institute of Radiation Medicine, Seoul National University Medical Research Center) ;
  • Kim, Hyungjin (Department of Radiology, Seoul National University College of Medicine, and Institute of Radiation Medicine, Seoul National University Medical Research Center) ;
  • Park, Chang Min (Department of Radiology, Seoul National University College of Medicine, and Institute of Radiation Medicine, Seoul National University Medical Research Center)
  • 변소현 (서울여자대학교 소프트웨어융합학과) ;
  • 정주립 (서울여자대학교 소프트웨어융합학과) ;
  • 홍헬렌 (서울여자대학교 소프트웨어융합학과) ;
  • 송용섭 (서울대학교병원 영상의학과) ;
  • 김형진 (서울대학교병원 영상의학과) ;
  • 박창민 (서울대학교병원 영상의학과)
  • Received : 2018.11.09
  • Accepted : 2018.11.30
  • Published : 2018.12.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we propose an automated method for the ground-glass nodule(GGN) classification using GGN-Net based on intensity, texture, and shape-enhanced images in chest CT images. First, we propose the utilization of image that enhances the intensity, texture, and shape information so that the input image includes the presence and size information of the solid component in GGN. Second, we propose GGN-Net which integrates and trains feature maps obtained from various input images through multiple convolution modules on the internal network. To evaluate the classification accuracy of the proposed method, we used 90 pure GGNs, 38 part-solid GGNs less than 5mm with solid component, and 23 part-solid GGNs larger than 5mm with solid component. To evaluate the effect of input image, various input image set is composed and classification results were compared. The results showed that the proposed method using the composition of intensity, texture and shape-enhanced images showed the best result with 82.75% accuracy.

본 논문에서는 흉부 CT 영상에서 결절의 밝기값, 재질 및 형상 증강 영상 기반의 GGN-Net을 이용해 간유리음영 결절 자동 분류 방법을 제안한다. 첫째, 입력 영상에 결절 내부의 고형 성분의 유무 및 크기 정보가 포함될 수 있도록 밝기값, 재질 및 형상 증강 영상의 활용을 제안한다. 둘째, 다양한 입력 영상을 여러 개의 컨볼루션 모듈을 통해 획득한 특징맵을 내부 네트워크에서 통합하여 훈련하는 GGN-Net를 제안한다. 제안 방법의 분류정확성 평가를 위해 순수 간유리음영 결절 90개와 고형 성분의 크기가 5mm 미만인 혼합 간유리음영 결절 38개, 5mm 이상 고형 성분의 크기를 가지는 혼합 간유리음영 결절 23개의 데이터를 사용하였으며, 입력 영상이 간유리음영 결절 분류 결과에 미치는 영향을 비교하기 위해 다양한 입력 영상을 구성하여 결과를 비교하였다. 실험 결과, 밝기값, 재질 및 형상 정보가 함께 고려된 입력 영상을 사용한 제안 방법이 정확도가 82.75%로 가장 좋은 결과를 보였다.

Keywords

References

  1. 장승훈, 등., "폐결절," 대한내과학회지, 제 79 권 부록 2호, 2010.
  2. C.I. Henschke, et al., "CT screening for lung cancer: frequency and significance of part-solid and non-solid nodules," Am.J.Roentgenol, Vol. 178, No. 5, pp. 1053-1057, 2002. https://doi.org/10.2214/ajr.178.5.1781053
  3. D.P. Naidich, et al., "Recommendations for the management of subsolid pulmonary nodules detected at CT: a statement from the Fleischner Society," Radiology, Vol266, No. 1, pp. 304-317, 2013. https://doi.org/10.1148/radiol.12120628
  4. H. MacMahon, et al, "Guidelines for management of incidental pulmonary nodules detected on CT images: from the Fleischner Society 2017," Radiology, Vol.284, No.1, pp. 228-243, 2017. https://doi.org/10.1148/radiol.2017161659
  5. M. Sergeeva, et al., "Classification of pulmonary nodules on computed tomography scans. Evaluation of the effectiveness of application of textural features extracted using wavelet transform of image," In Proveedings of the 18th Conference of Open Innovations Association FRUCT, pp. 291-299, 2016.
  6. S.G. Armato, et al., "LUNGx Challenge for computerized lung nodule classification," Journal of Medical Imaging, vol. 3, no. 4, pp. 044506-044506, 2016. https://doi.org/10.1117/1.JMI.3.4.044506
  7. D. Kumar, et al., "Lung nodule classification using deep features in CT images," In Computer and Robot vision(CRV), 2015 12th Conference on. IEEE, pp. 133-138, 2015.
  8. S.Y. Lee, et al., "Classification of Ground-Glass Opacity Nodules with Small Solid Components using Multiview Images and Texture Analysis in Chest CT Images", Journal of Korea Multimedia Society, Vol.20, No.7, pp. 994-1003, 2017. https://doi.org/10.9717/KMMS.2017.20.7.994
  9. S.Y. Lee, et al., "Multi-class Classification of Ground-glass Nodules using 2.5-dimensional Multiview-based Features and Data Augmentation" KIISE Transactions on Computing Practices, Vol.24, No.10, pp. 527-533, 2018. https://doi.org/10.5626/KTCP.2018.24.10.527
  10. N. Tajbakhsh, et al., "Comparing two classes of end-to-end machine-learning models in lung nodule detection and classification: MTANNs vs. CNNs", Pattern recognition, Vol.63, pp. 476-486, 2017. https://doi.org/10.1016/j.patcog.2016.09.029
  11. K. Liu, et al., "Multiview Convolutional Neural Networks for lung nodule classification," International Journal of Imaging Systems and Technology, Vol.27, No.1, pp. 12-22, 2017. https://doi.org/10.1002/ima.22206
  12. W. Shen, et al., "Multi-crop convolutional neural networks for lung nodule malignancy suspiciousness classification," Pattern Recognition, Vol.61, pp.663-673, 2017. https://doi.org/10.1016/j.patcog.2016.05.029
  13. R. Dey, et al., "Diagnostic classification of lung nodules using 3D neural networks" In Biomedical Imaging (ISBI 2018), IEEE 15th International Symposium on, pp. 774-778, 2018.
  14. B. Xu, et al., "Empirical Evaluation of Rectified Activations in Convolutional Network, arXiv preprint ArXiv:1505.00853.