• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.5
• /
• pp.323-330
• /
• 2023

The development of an analysis model that reflects the microstructure characteristics of polyvinyl alcohol (PVA) fiber-reinforced cementitious composites, which have a highly complex microstructure, enables synergy between efficient material design and real experiments. PVA fiber orientations are an important factor that influences the mechanical behavior of PVA fiber-reinforced cementitious composites. Owing to the difficulty in distinguishing the gray level value obtained from micro-CT images of PVA fibers from adjacent phases, fiber segmentation is time-consuming work. In this study, a micro-CT test with a voxel size of 0.65 ㎛³ was performed to investigate the three-dimensional distribution of fibers. To segment the fibers and generate training data, histogram, morphology, and gradient-based phase-segmentation methods were used. A U-net model was proposed to segment fibers from micro-CT images of PVA fiber-reinforced cementitious composites. Data augmentation was applied to increase the accuracy of the training, using a total of 1024 images as training data. The performance of the model was evaluated using accuracy, precision, recall, and F1 score. The trained model achieved a high fiber segmentation performance and efficiency, and the approach can be applied to other specimens as well.

• Journal of Internet Computing and Services
• /
• v.24 no.4
• /
• pp.85-92
• /
• 2023

Deep learning applications are increasingly being leveraged for disease detection tasks in medical imaging modalities such as X-ray, Computed Tomography (CT), and Magnetic Resonance Imaging (MRI). Most data-centric deep learning challenges necessitate the use of supervised learning methodologies to attain high accuracy and to facilitate performance evaluation through comparison with the ground truth. Supervised learning mandates a substantial amount of image and label sets, however, procuring an adequate volume of medical imaging data for training is a formidable task. Various data augmentation strategies can mitigate the underfitting issue inherent in supervised learning-based models that are trained on limited medical image and label sets. This research investigates the enhancement of a deep learning-based rib fracture segmentation model and the efficacy of data augmentation techniques such as left-right flipping, rotation, and scaling. Augmented dataset with L/R flipping and rotations(30°, 60°) increased model performance, however, dataset with rotation(90°) and ⨯0.5 rescaling decreased model performance. This indicates the usage of appropriate data augmentation methods depending on datasets and tasks.

• The Journal of Engineering Geology
• /
• v.24 no.3
• /
• pp.353-362
• /
• 2014

Pore structures in porous rock play an important role in hydraulic & mechanical behaviour of rock. Porosity, size distribution and orientation of pores represent the characteristics of pore structures of porous rock. While effective porosity can be measured easily by conventional experiment, pore size distribution is hard to be quantified due to the lack of corresponding experiment. We assessed pore size distribution of Berea sandstone using X-ray CT image based analysis combined with associated images processing, i.e., image filtering, binarization and skeletonization subsequently followed by the assessment of local thickness and star chord length. The aim of this study is to propose a new and effective way to evaluate pore structures of porous rock using X-ray CT based analysis for pore size distribution.

• Journal of the Korean Geotechnical Society
• /
• v.39 no.8
• /
• pp.17-28
• /
• 2023

Laboratory-scale hydraulic fracturing experiments were conducted on granite specimens at various viscosities and injection rates of the fracturing fluid. A series of cross-sectional computed tomography (CT) images of fractured specimens was obtained via a three-dimensional X-ray CT imaging method. Pixel-level fracture segmentation of the CT images was conducted using a convolutional neural network (CNN)-based Nested U-Net model structure. Compared with traditional image processing methods, the CNN-based model showed a better performance in the extraction of thin and complex fractures. These extracted fractures extracted were reconstructed in three dimensions and morphologically analyzed based on their fracture volume, aperture, tortuosity, and surface roughness. The fracture volume and aperture increased with the increase in viscosity of the fracturing fluid, while the tortuosity and roughness of the fracture surface decreased. The findings also confirmed the anisotropic tortuosity and roughness of the fracture surface. In this study, a CNN-based model was used to perform accurate fracture segmentation, and quantitative analysis of hydraulic stimulated fractures was conducted successfully.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1999.04a
• /
• pp.105-107
• /
• 1999

본 연구의 목적은 CT장비로부터 얻어지는 단면 영상을 이용하여 재구성한 3차원 Voxel 정보를 기반으로 의료 시술 중 위험도가 높으며 장기간의 수술 훈련이 필요한 수술인 척추경나사 삽입술에 대한 모의 시술기를 개발하는 것이다. 모의 시술기의 입력은 환자의 환부에 대한 CT와 모의 시술을 해보고자 하는 의사 (사용자)의 입력 (경나사의 진입 위치와 각도)이 되며 출력은 의사들이 시술장에서 받을 수 있는 유일한 방법인 Voxel데이터로부터 재생성된 X-Ray이미지, 혹은 C-Arm의 동영상이며, 최종 결과 출력은 나사못이 삽입된 재구성 CT 이미지들과 3차원 정보를 볼 수 있는 Image Based Rendering의 Image data set이 된다. 본 연구에서는 각 시각화 부분의 특성을 고려하여 direct volume projection, surface modeling, 그리고 최근 많은 관심을 받고 있는 Image Based Rendering 기법을 intergrate하여 사용하였으며 각 시각화 모듈의 초고속 정보 통신망에서의 정보 교환에 대한 방법론에 대해 다루고 있다.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2009.06a
• /
• pp.1023-1024
• /
• 2009

• Progress in Medical Physics
• /
• v.24 no.4
• /
• pp.265-270
• /
• 2013

Correct target volume delineation is an important part of radiosurgery treatment planning process. We designed head phantom and performed target delineation to evaluate the volume differences due to radiosurgery treatment planning systems and image acquisition system, CT/MR. Delineated mean target volume from CT scan images was $2.23{\pm}0.08cm^3$ on BrainSCAN (NOVALS), $2.13{\pm}0.07cm^3$ on Leksell gamma plan (Gamma Knife) and $2.24{\pm}0.10cm^3$ on Multi plan (Cyber Knife). For MR images, $2.08{\pm}0.06cm^3$ on BrainSCAN, $1.94{\pm}0.05cm^3$ on Leksell gamma plan and $2.15{\pm}0.06cm^3$ on Multi plan. As a result, Differences of delineated mean target volume due to radiotherapy planning system was 3% to 6%. And overall mean target volume from CT scan images was 6.36% larger than those of MR scan images.

• Conservation Science in Museum
• /
• v.25
• /
• pp.63-84
• /
• 2021

In order to investigate the diverse physicochemical changes that occurred in traditional Korean pottery during its production, including before and after firing, this study produced six replicas of a celadon maebyeong inlaid with cloud-and-crane designs, respectively corresponding to the process of shaping, carving, inlaying designs, first firing, glazing and second firing, respectively. It then conducted a scientific study of these six replicas and analyzed their images through high-resolution three-dimensional transmission imaging. The materials used for the replicas show different mineral phases and even colors depending on the components of each material. For example, black inlay with a high content of iron oxide (Fe₂O₃) shows dark colors and white inlay with a high alumina (Al₂O₃) content appears white. Physicochemical properties such as chromaticity and magnetic susceptibility and major components of the replicas were confirmed by the differences in the density in the computed tomography (CT) images. The characteristics of fired products such as fine structure, absorption ratio, apparent porosity, and other characteristics of the major mineral components were identified by the presence of pores and the formation of cracks inside the replicas in the image analysis.

• Proceedings of the KSMRM Conference
• /
• 2003.10a
• /
• pp.78-78
• /
• 2003

목적： 이전에는 손쉽게 구할 수 있는 표준데이터를 가지고 visual human body model을 형성하였다. 주로 팬텀이나, 외국인의 데이터를 가지고 만든 것이기 때문에 우리가 실제 실험에 쓰려면 큰 차이가 있었다. 그래서 본 연구에서는 실제 우리나라 사람 중 동일 인물의 MR와 CT 이미지를 가지고 인체 모델을 만들고자 하였다. 그러기 위해서 먼저 인체의 MR, CT영상에 대한 특징을 분석해야 했고, 이것을 바탕으로 영상 분할(Image Segmentation)을 하였다. 인체 부위에 따라 영상 분할 방법도 그 차이가 있음을 알 수 있었다.

• Journal of the Korea Computer Graphics Society
• /
• v.11 no.2
• /
• pp.26-33
• /
• 2005

의료기기들 중 기능 영상을 보기 위해 이용되는 PET 장치에서 획득된 결과 영상은 선명하지 않기 때문에, 해부학적 구조와 기능 영상을 동시에 보기 위해서는 선명한 영상을 제공하는 CT 와 PET 장치와 하나로 통합하여 영상을 획득하게 되었다. 그래서 한번의 촬영으로 PET/CT 영상을 얻을 수 있게 된 것이다. 서로 다른 특성을 갖는 이미지를 융합하게 되면 보다 정확한 진단을 내리는데 많은 도움을 준다. 본 논문은 CT 영상에서 폐 영역을 반 자동(Semi-Auto)으로 분리한 후 PET 영상에 자동으로 융합하는 방법을 제안한다. 반 자동 폐 영역 분할을 위해 1 차원 신호 처리 기법과 Seeded Region Growing 기법을 사용한다. 수행된 폐 분리 결과는 몸의 해부학적 구조를 보기 위해 사용되는 CT 영상에서 추출한 폐 영역을 기능을 보기 위한 PET 영상에 퓨전 함으로서 진단 전문가가 보다 정확한 진단을 하는데 도움이 될 것이다. 또한 이러한 기능을 쉽게 구현하고 사용할 수 있도록 시각 프로그래밍 기법을 접목하였다.