• Investigative Magnetic Resonance Imaging
• /
• v.22 no.4
• /
• pp.229-239
• /
• 2018

Purpose: To assess the diagnostic performance in detecting rotator cuff tears at 3T of non-arthrographic shoulder magnetic resonance imaging (MRI) using 3D isotropic turbo spin-echo (TSE-SPACE) sequence as compared with 2D sequences. Materials and Methods: Seventy-four patients who were arthroscopically confirmed to have underwent non-arthrographic shoulder MRI with 2D sequences and TSE-SPACE were included. Three independent readers retrospectively scored supraspinatus and infraspinatus tendon (SST-IST) and subscapularis tendon (SCT) tears on 2D sequences and TSE-SPACE. Results: The mean sensitivity, specificity, and accuracy of the three readers were 95%, 100%, and 95% on TSE-SPACE and 99%, 93%, and 98% on 2D sequences for detecting SST-IST tears, respectively, whereas those were 87%, 49%, and 68% on TSESPACE and 88%, 66%, and 77% on 2D sequences for detecting SCT tears, respectively. There was no statistical difference between the two sequences, except for in the specificity of one reader for detecting SCT tears. The mean AUCs of the three readers on TSE-SPACE and 2D sequences were 0.96 and 0.98 for detecting SST-IST tears, respectively, which were not significantly different, while those were 0.71 and 0.82 for detecting SCT tears, respectively, which were significantly different (P < 0.05). Conclusion: TSE-SPACE may have accuracy and reliability comparable to conventional 2D sequences for SST-IST tears at non-arthrographic 3T shoulder MRI, whereas TSE-SPACE was less reliable than conventional 2D sequences for detecting SCT tears.

• Journal of the Korean Society of Radiology
• /
• v.10 no.6
• /
• pp.403-409
• /
• 2016

MRI scan is a useful method in the diagnosis of musculoskeletal excellent contrast of the organization. Depending on the patient's musculoskeletal examinations state the type of aids provided the aid is used there is also challenging as well as the costs do not vary. This study was produced by the use of 3D printing technology, an MRI aids. Aids in the production process, then through 3D modeling and then convert stl files using (3D MAX.2014, Fusion360) slicing programs (Cubicreater 2.1ver., Cura 15.4ver) converted to G-code printed on the FDM scheme (Cubicon Style, output was MICRO MAKE). Output is, but in the FDM to evaluate the SNR on the MRI images were compared to the test is the case before use, and then to produce a Water Phantom case of a PLA, ABS, a TPU thickness 3mm, using aids before, It was evaluated in a clinical image after qualitatively. Obtaining an image of SNR Warter Phantom appeared to have been evaluated as T1 NON $123.778{\pm}28.492$, PLA $123.522{\pm}28.373$, ABS $124.461{\pm}25.716$, TPU $124.843{\pm}27.272$. T2 NON $127.421{\pm}26.949$, was rated as PLA $124.501{\pm}27.768$, ABS $128.663{\pm}26.549$, TPU $130.171{\pm}25.998$. The results did not show statistically significant differences. The use of assistive devices before and after images Clinical evaluation method palliative $3.20{\pm}0.88$, $3.95{\pm}0.76$ after using the aids used to aid improved the quality of the image. Production of the auxiliary mechanism using a future 3D printing is expected are thought to be used clinically, it can be an aid making safe and comfortable than the inspection of the patient is an alternative to improve the problems of the aids used in the conventional do.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.04a
• /
• pp.336-339
• /
• 2012

VTK is a free but professional development platform for images three-dimensional (3D) reconstruction and processing. It is powerful, open-source, and users can customize their own needs by self-development of great flexibility. To give the doctors more and detailed information by simulate dissection to the 3-D brain MRI image after reconstruction. A Visualization System (VS) is proposed to achieve 3D brain reconstruction and virtual dissection functions. Based on the free VTK visualization development platform and Visual Studio 2010 IDE development tools, through C++ language, using real people's MRI brain dataset, we realized the images 3D reconstruction and also its applications and extensions correspondingly. The display effect of the reconstructed 3D image is well and intuitive. With the related operations such as measurement, virtual dissection and so on, the good results we desired could be achieved.

• Journal of Korea Multimedia Society
• /
• v.23 no.8
• /
• pp.940-952
• /
• 2020

The Alzheimer's disease (AD) is a neurodegenerative disease commonly found in the elderly individuals. It is one of the most common forms of dementia; patients with AD suffer from a degradation of cognitive abilities over time. To correctly diagnose AD, compuated-aided system equipped with automatic classification algorithm is of great importance. In this paper, we propose a novel deep learning based classification algorithm that takes advantage of MRI biomarker images including brain areas of hippocampus and cerebrospinal fluid for the purpose of improving the AD classification performance. In particular, we develop a new approach that effectively applies MRI biomarker patch images as input to 3D Deep Convolution Neural Network. To integrate multiple classification results from multiple biomarker patch images, we proposed the effective confidence score fusion that combine classification scores generated from soft-max layer. Experimental results show that AD classification performance can be considerably enhanced by using our proposed approach. Compared to the conventional AD classification approach relying on entire MRI input, our proposed method can improve AD classification performance of up to 10.57% thanks to using biomarker patch images. Moreover, the proposed method can attain better or comparable AD classification performances, compared to state-of-the-art methods.

• Journal of Korean Foot and Ankle Society
• /
• v.7 no.2
• /
• pp.208-217
• /
• 2003

Introduction: Soft-tissue impingement syndrome is now increasingly recognized as a significant cause of the chronic ankle pain. As a method to detect soft-tissue ankle impingement, a characteristic history and physical examination, routine MR imaging, and direct MR arthrography were used. The efficacy of routine MR imaging has been controversial for usefulness because of low sensitivity and specificity. Direct MR artrhography was recommaned for diagnosis because of the highest sensitivity, specificity and accuracy, but it requires an invasive procedure. The purpose of this study is to investigate the diagnostic accuracy of Fat suppressed, contrast enhanced, three-dimensional fast gradient recalled acquisition in the steady state with rediofrequency spoiling magnetic resonance imaging(CE 3D-FSPGR MRI) and to evaluate the clinical outcome of the arthroscopic treatment in assessing soft-tissue impingement associated with trauma of the ankle. Materials and Methods: We reviewed 38 patients who had arthroscopic evaluations and preoperative magnetic resonance imaging studies(3D-FSPGR MRI) for post-traumatic chronic ankle pain between January 2000 and August 2002. Among them, 24 patients had osteochondral lesion, lateral instability, loose body, malunion of lateral malleoli, and peroneal tendon dislocation. The patient group consisted of 23 men and 15 women with the average age of 34 years(16-81 years). The mean time interval from the initial trauma to the operation was 15.5 months(3 to 40 months), The mean follow-up duration of the assessment was 15.6months(12-48 months). MRI was simultaneously reviewed by two radiologists blinded to the clinical diagnosis. The sensitivity, specificity and accuracy of MRI was obtained from radiologic and arthroscopic finding. Arthroscopic debridement and additional operation for associated disease were performed. We used a standard protocol to evaluate patients before the operation and at follow-up which includes American Orthopedic Foot and Ankle Society Ankle-Hindfoot Score. Results: For the assessment of the synovitis and soft tissue impingement, fat suppressed CE 3D-FSPGR MR imaging had the sensitivity of 91.9%, the specificity of 84.4 and the accuracy of 87.5%. AOFAS Ankle-Hindfoot Score of preoperative state was 69.2, and the mean score of the last follow-up was 89.1. These were assessed as having 50% excellent(90-100) and 50% good(75-89). The presence of other associated disease didn't show the statistically significant difference(>0.05). Conclusion: Fat suppressed CE 3D-FSPGR MR imaging is useful method comparable to MR arthrography for diagnosis of synovitis or soft-tissue impingement, and arthroscopic debridement results in good clinical outcome.

• Journal of Biomedical Engineering Research
• /
• v.42 no.4
• /
• pp.143-149
• /
• 2021

In order to predict and prevent the disease of the abdominal aorta, which is the largest artery in the human body and the most common aneurysm, the normal arterial blood flow operation should be considered. To this end, we are trying to solve problems that may arise in the future by executing FSI based on the data obtained from 4D flow MRI. However, to match the similarity between the 4D flow MRI flow and the FSI flow, correlation was used in previous papers, but the correlation did not show the degree of agreement. Therefore, in this paper, we analyzed the correlation between the 4D flow MRI flow velocity of the human abdominal aorta and the two-way FSI flow velocity in which the three physical properties used for the aortic FSI were added to the CT abdominal aorta 3D model and the interclass correlation coefficient. As a result, the physical property M2 showed the highest similarity in correlation and intraclass correlation coefficient, and this property is intended to be helpful in the future study of the abdominal aortic two-way FSI flow rate.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.145-146
• /
• 2017

본 논문에서는 남성의 하반신을 촬영한 MRI 영상으로부터 전립선을 분할하는 알고리즘을 제안한다. 우선 3 차원 입체 영상을 학습하기 위해 3D 컨볼루션 계층(convolutional layer) 및 3D 풀링 계층(pooling layer)에 기반한 네트워크를 제안한다. 다음으로 네트워크의 최후단에 해당하는 전연결 계층(fully connected layer)의 강인한 학습을 돕는 잡음 계층을 제안한다. 잡음 계층은 네트워크의 학습 파라미터 혹은 출력 영상에 가우시안 잡음를 더함으로써 드롭 아웃과 같이 훈련 영상에 대한 과적합(overfitting)을 막고 테스트 영상에 강인한 네트워크의 학습을 돕는다. 마지막으로 실험을 통해 제안하는 기법이 기존 기법에 비해 우수한 분할 성능을 보임을 확인한다.

• Journal of Korea Society of Industrial Information Systems
• /
• v.20 no.3
• /
• pp.81-86
• /
• 2015

We fabricated a 3D knee joint model through the imaging processing. The 3D shape information is different depends on specific conditions when the shape of real knee joint is extracted from CT/MRI sliced images. The two types of joint models were fabricated by using 3D printer in order to analysis of joint movement by slight difference of 3D shape information. The compressive force experiments were performed by using knee joint model. As the results, the compressive forces were changed with respect to the difference of geometry. Consequently, feasibility test should be performed before developing biomimetic bioreactor.

• Bulletin of the Korean Chemical Society
• /
• v.32 no.3
• /
• pp.1022-1026
• /
• 2011

The syntheses of DTPA-bis(amide) conjugates of phosphonated cyclohexane moieties (5a-d) and their Gd(III) complexes of the type $[Gd(L)(H_2O)]{\cdot}nH_2O$ (6a-d; L = 5a-d) are described. All new compounds have been characterized by microanalysis and spectroscopic techniques. High $r_1$ relaxivities of aqueous solutions of 6a-d are observed to be in the range of $10.7-18.3\;mM^{-1}sec^{-1}$, which compare much better than that of $Omniscan^{(R)}$ ($r_1=3.90\;mM^{-1}sec^{-1}$).

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.416-418
• /
• 2022

Glioma is a type of brain tumor that occurs in glial cells and is classified into two types: high hrade hlioma with a poor prognosis and low grade glioma. Magnetic resonance imaging (MRI) as a non-invasive method is widely used in glioma diagnosis research. Studies to obtain complementary information by combining multiple modalities to overcome the incomplete information limitation of single modality are being conducted. In this study, we developed a 3D CNN-based model that applied input-level fusion to MRI of four modalities (T1, T1Gd, T2, T2-FLAIR). The trained model showed classification performance of 0.8926 accuracy, 0.9688 sensitivity, 0.6400 specificity, and 0.9467 AUC on the validation data. Through this, it was confirmed that the grade of glioma was effectively classified by learning the internal relationship between various modalities.