• Journal of the Korean Society of Radiology
• /
• v.18 no.5
• /
• pp.541-549
• /
• 2024

The various structures of artificial neural networks, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have been extensively studied and served as the backbone of numerous models. Among these, a transformer architecture has demonstrated its potential for natural language processing and become a subject of in-depth research. Currently, the techniques can be adapted for image processing through the modifications of its internal structure, leading to the development of Vision transformer (ViT) models. The ViTs have shown high accuracy and performance with large data-sets. This study aims to develop a ViT-based model for detecting pneumonia using chest X-ray images and quantitatively evaluate its performance. The various architectures of the ViT-based model were constructed by varying the number of encoder blocks, and different patch sizes were applied for network training. Also, the performance of the ViT-based model was compared to the CNN-based models, such as VGGNet, GoogLeNet, and ResNet. The results showed that the traninig efficiency and accuracy of the ViT-based model depended on the number of encoder blocks and the patch size, and the F1 scores of the ViT-based model ranged from 0.875 to 0.919. The training effeciency of the ViT-based model with a large patch size was superior to the CNN-based models, and the pneumonia detection accuracy of the ViT-based model was higher than that of the VGGNet. In conclusion, the ViT-based model can be potentially used for pneumonia detection using chest X-ray images, and the clinical availability of the ViT-based model would be improved by this study.

• Investigative Magnetic Resonance Imaging
• /
• v.17 no.2
• /
• pp.133-143
• /
• 2013

Purpose : To evaluate and compare the accuracy of magnetic resonance imaging (MRI) and ultrasound (US) for detection and estimation of invasion depth of colorectal carcinoma (CRC) by correlation with histopathologic findings in vitro, and to find out the best MR pulse sequence for accurate delineation of tumor from surrounding normal tissue. Materials and Methods: Resected specimens of CRC from 45 patients were examined about tumor detectability and invasion depth of US using high frequency (5-17 MHz) linear transducer in a tube filled with normal saline and MRI in a 8-channel quadrate head coil. The institutional review board approved this study and informed consent was waived. MRI with seven pulse sequences of in- and out-of-phases gradient echo T1 weighted images, fast spin echo T2 weighted image and its fat suppression image, fast imaging employing steady-state acquisition (FIESTA) and its fat suppression image, and diffusion weighted image (DWI) were performed. In each case, both imaging findings of MRI and US were evaluated independently for detection and estimation of invasion depth of tumor by consensus of two radiologists and were compared about diagnostic accuracy according to the histopathologic findings as reference standard. Seven MR pulse sequences were evaluated on the point of accurate delineation of tumor from surrounding normal tissue in each specimen. Results: In specimens of CRC, both imaging modalities of MRI (91.1%) and US (86.7%) showed relatively high diagnostic accuracy to detect tumor and evaluate invasion depth of tumor. In early CRC, diagnostic accuracy of US was 87.5% and that of MRI was 75.0%. There was no statistically significant difference between two imaging modalities (p > 0.05). The best pulse sequence among seven MR sequences for accurate delineation of tumor from surrounding normal tissue in each specimen of CRC was fast spin echo T2 weighted image. Conclusion: MRI and US show relatively high diagnostic accuracy to detect tumor and evaluate invasion depth of resected specimen of CRC. The most excellent pulse sequence of MRI for accurate delineation of tumor from surrounding normal tissue in CRC is fast spin echo T2 weighted image.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.22 no.6
• /
• pp.643-653
• /
• 2020

The concrete lining inspection by inspector after traffic control poses various problems such as congestion caused by traffic control and social loss costs resulting from it, and risks of traffic accidents and safety accidents for inspectors et al. To solve the problems, the concrete lining inspection has been carried out using automatic inspection equipment and image analysis that can be objectively and quantitatively investigated in overseas. In this study, to solve the problems of concrete lining inspection by inspector and to review improvement plan for inspection, inspection was carried out using automatic inspection equipment for ◯◯ tunnel that precision safety diagnosis has been conducted in 2019. Analysis was carried out for both inspection results. Automatic inspection equipment investigated defects in concrete lining such as cracks more accurately than precision safety diagnosis.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.18 no.1
• /
• pp.166-173
• /
• 2014

A noncontact nondestructive testing (NDT) method is proposed to detect the damage of pipeline structures and to identify the location of the damage. To achieve this goal, a scanning laser source actuation technique is utilized to generate a guided wave and scans a specific area to find damage location more precisely. The ND: YAG pulsed laser is used to generate Lamb wave and a piezoelectric sensor is installed to measure the structural responses. The measured responses are analyzed using three dimensional Fourier transformation (3DFT). The damage-sensitive features are extracted by wavenumber filtering based on the 3D FT. Then, flaw imaging techniques of a pipeline structures is conducted using the damage-sensitive features. Finally, the pipes with notches are investigated to verify the effectiveness and the robustness of the proposed NDT approach.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.490-492
• /
• 2016

Pancreatic ultrasound imaging is used to diagnose pancreatic hyperplasia, pancreatic steatosis, pancreatic cancer and the like. If the diagnosis of pancreatic steatosis is pancreatic parenchyma echo shades splashes spleen than in the pancreas ultrasound it determines that the fat is deposited. And research on ultrasound imaging of pancreatic cancer but is actively conducted research studies on pancreatic steatosis is insufficient In addition, pancreatic steatosis is often an error in accordance with the diagnostic criteria are vague and subjective diagnosis of the artisan. This study was a quantitative analysis using the feature value extracting a feature of an image extracted by applying a parameter to the algorithm GLCM image of the normal and pancreatic fat. Setting a region of interest ($5{\times}5pixel$) in the mild 89 case, moderate 89 case, severe 89 case, total image 267 case using GLCM algorithm, and using the Autocorrelation, Sum average, Sum of squares, Sum varience 4 kinds parameter in each image It was analyzed.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2004.05a
• /
• pp.13-18
• /
• 2004

최근 몇 년간 방사선 의학진단과 관련된 연구가 한층 높아진 가운데 유방암은 여성의 암 중에서 1위를 차지하고 조기에 진단하고 치료하기 위한 국가적인 노력이 필요한 시점이다. 이렇듯 여성들의 유방암 발생빈도수가 급증하면서 대두 되고 있는 것이 조기 진단방법인 Mammography와 초음파 진단이며 그로인하여 발생하는 오진률 역시 많은 연구가 진행 되고 있다. 먼저 Mammography 및 초음파 진단의 문제점 보면 첫째 촬영과정에서의 오차, 둘째 영상의 선명도 ,셋째 전문의의 판독에 대한오차, 넷째 의사의 경험으로 진단함으로 표준화가 존재하지 않는다는 공통적인 문제점을 가지고 있다. 본 연구에서는 CAD 시스템의 프레임웍 및 요소 기술을 제시하여 의사의 진단을 보조적 수행이 보다 수월하도록 하고자 한다. 본 연구에서는 CAD시스템의 기능은 Detection기능(Image enhancement, Morphology, segment detection)과 Diagnosis기능( Neural Natwork등을 이용하여 증상을 판단)이다. 또한 과거 자료를 이용한 변이 및 변화를 예측함으로써 향후 있을 위험요소에 대비가 가능한 모듈과 전문의사가 대화형으로 빠르게 진단지식을 구축할 수 있는 지능형, 대화형 온라인 진단기능을 추가함으로써 외국의 CAD시스템과는 많은 차이가 있다고 볼 수 있다.

• Proceedings of the KSMRM Conference
• /
• 2002.11a
• /
• pp.99-99
• /
• 2002

목적： 인체에 전류를 주입하면 체내의 생체조직의 임피던스 분포에 따라서 전류밀도 분포가 결정된다. 이러한 전류밀도 분포에 대한 정보는 전기임피던스 단층촬영법과 유방암 진단, 체내 온도 분포의 영상화, 전기자극에 의한 체내 전류 경로의 시각화에 대한 연구에 응용될 수 있다. 한편 이러한 전류밀도 분포는 전류주입 자기공명영상기법에 의해 영상화할 수 있으며, 본 논문은 3차원 팬텀 내부의 전류밀도 분포를 영상화하는 전류주입 자기공명영상기법의 실험결과를 기술한다.

• Proceedings of the Optical Society of Korea Conference
• /
• 2003.07a
• /
• pp.76-77
• /
• 2003

Digital Radiograpy ＆ Fluoroscopy(DRF 또는 DR 또는 DF)는 cone beam을 이용하여 인체를 투과한 X선을 영상증배관(Image Intensifying Tube： IIT)을 통하여 가시광선으로 변환시킨 후 영상을 카메라로 보내고, 이곳에서 발생한 영상정보를 디지털로 처리하여 모니터를 통해 눈에 보이는 영상으로 만드는 방법으로 IIT에 기초한 디지털 방사선 촬영술이라고도 한다. DF 방법은 즉시 영상 표시와 진단이 가능하기 때문에 즉시성이 요구되는 심장이나 두복부 등의 순환기 분야에서 DSA(Digital Subtraction Angiography) 장비로 이용되고 있고, 순환기뿐만 아니라 위를 중심으로 한 소화관(식도, 위, 소장, 대장, 직장)의 분야에서 적용 가능하다. (중략)

• Journal of the Korean Society of Radiology
• /
• v.81 no.1
• /
• pp.58-69
• /
• 2020

In this study, we evaluated the technical characteristics and usefulness of diffusion-weighted magnetic resonance imaging for discrimination between benign and malignant vertebral fractures, for detection and differentiation of multiple myeloma or metastases, and for response monitoring in malignant vertebral lesions after anticancer drug therapy or radiation therapy.

• Journal of the Korean Society of Radiology
• /
• v.82 no.6
• /
• pp.1505-1523
• /
• 2021

Purpose Although chest CT has been discussed as a first-line test for coronavirus disease 2019 (COVID-19), little research has explored the implications of CT exposure in the population. To review chest CT protocols and radiation doses in COVID-19 publications and explore the number needed to diagnose (NND) and the number needed to predict (NNP) if CT is used as a first-line test. Materials and Methods We searched nine highly cited radiology journals to identify studies discussing the CT-based diagnosis of COVID-19 pneumonia. Study-level information on the CT protocol and radiation dose was collected, and the doses were compared with each national diagnostic reference level (DRL). The NND and NNP, which depends on the test positive rate (TPR), were calculated, given a CT sensitivity of 94% (95% confidence interval [CI]: 91%-96%) and specificity of 37% (95% CI: 26%-50%), and applied to the early outbreak in Wuhan, New York, and Italy. Results From 86 studies, the CT protocol and radiation dose were reported in 81 (94.2%) and 17 studies (19.8%), respectively. Low-dose chest CT was used more than twice as often as standard-dose chest CT (39.5% vs.18.6%), while the remaining studies (44.2%) did not provide relevant information. The radiation doses were lower than the national DRLs in 15 of the 17 studies (88.2%) that reported doses. The NND was 3.2 scans (95% CI: 2.2-6.0). The NNPs at TPRs of 50%, 25%, 10%, and 5% were 2.2, 3.6, 8.0, 15.5 scans, respectively. In Wuhan, 35418 (TPR, 58%; 95% CI: 27710-56755) to 44840 (TPR, 38%; 95% CI: 35161-68164) individuals were estimated to have undergone CT examinations to diagnose 17365 patients. During the early surge in New York and Italy, daily NNDs changed up to 5.4 and 10.9 times, respectively, within 10 weeks. Conclusion Low-dose CT protocols were described in less than half of COVID-19 publications, and radiation doses were frequently lacking. The number of populations involved in a first-line diagnostic CT test could vary dynamically according to daily TPR; therefore, caution is required in future planning.