A5014835747- Radiomics and Machine Learning in Medical Imaging
- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Radiation Dose and Imaging
- Lung Cancer Diagnosis and Treatment
- AI in cancer detection
- Digital Radiography and Breast Imaging
- Advanced MRI Techniques and Applications
- MRI in cancer diagnosis
- Medical Image Segmentation Techniques
- Chronic Obstructive Pulmonary Disease (COPD) Research
- Advanced Data Compression Techniques
- Atomic and Subatomic Physics Research
- Colorectal Cancer Screening and Detection
- Medical Imaging and Analysis
- Image and Signal Denoising Methods
- Gastric Cancer Management and Outcomes
- Gastrointestinal Bleeding Diagnosis and Treatment
- Pleural and Pulmonary Diseases
- Inhalation and Respiratory Drug Delivery
- Artificial Intelligence in Healthcare
- Radiology practices and education
- Retinal Imaging and Analysis
- Advanced Radiotherapy Techniques
- Computer Graphics and Visualization Techniques
Seoul National University
2015-2025
Seoul National University Hospital
2014-2025
Advanced Institute of Convergence Technology
2015-2025
Paulsson (United States)
2025
Seoul Institute
2021-2022
Chung-Ang University Hospital
2021
New Generation University College
2002-2019
Seoul National University of Science and Technology
2019
Kyungpook National University Hospital
2017
Armed Forces Capital Hospital
2017
To perform a radiogenomic analysis of women with breast cancer to study the multiscale relationships among quantitative computer vision-extracted dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging phenotypes, early metastasis, and long noncoding RNA (lncRNA) expression determined by means high-resolution next-generation sequencing.In this institutional review board-approved study, an automated image platform extracted 47 computational features from DCE MR data in...
Purpose To evaluate the potential of xenon ventilation computed tomography (CT) in quantitative and visual analysis chronic obstructive pulmonary disease (COPD). Materials Methods This study was approved by institutional review board. After informed consent obtained, 32 patients with COPD underwent CT performed before administration xenon, two-phase wash-in (WI) wash-out (WO) periods, function testing (PFT). For analysis, results PFT were compared attenuation parameters from prexenon images...
OBJECTIVE To assess the feasibility of applying a deep learning-based denoising technique to coronary CT angiography (CCTA) along with iterative reconstruction for additional noise reduction. MATERIALS AND METHODS We retrospectively enrolled 82 consecutive patients (male:female = 60:22; mean age, 67.0 ± 10.8 years) who had undergone both CCTA and invasive artery from March 2017 June 2018. All included underwent (ADMIRE level 3, Siemens Healthineers). developed learning based (ClariCT.AI,...
Purpose: Reducing the patient dose while maintaining diagnostic image quality during CT exams is subject of a growing number studies, in which simulations reduced-dose with data have been used as an effective technique when exploring potential various reduction techniques. Difficulties accessing raw sinogram data, however, restricted use this to limited institutions. Here, we present novel simulation provides realistic low-dose images without requirement data. Methods: Two key...
Objective: This study evaluated the applicability of classical method, height loss ratio (HLR), for identifying major acute compression fractures in clinical practice and compared its performance with deep learning (DL)-based VCF detection methods. Additionally, it examined whether combining HLR DL approaches could enhance performance, exploring potential integration methodologies. Methods: End-to-End Detection (EEVD), Two-Stage Segmentation (TSVD_SD), Classification (TSVD_DC). The models...
Recent advancements in deep learning have significantly improved medical image segmentation. However, the generalization performance and potential risks of data-driven models remain insufficiently validated. Specifically, unrealistic segmentation predictions deviating from actual anatomical structures, known as a Seg-Hallucination, often occur learning-based models. The Seg-Hallucinations can result erroneous quantitative analyses distort critical imaging biomarker information, yet effective...
Correct pulmonary nodule volumetry and categorization is paramount for accurate diagnosis in lung cancer screening programs. CT scanners with slice thicknesses of multiple millimetres are still common worldwide, thickness has an adverse effect on the accuracy volumetry. We propose a deep learning based super-resolution technique to generate thin-slice images from thick-slice images. Analysis was performed using commercially available AI-based software. The increased 72.7 percent 94.5 when...
The purpose of this study was to develop an automated scheme facilitate detection localized ground-glass opacity (GGO) in the lung at computed tomography (CT). Institutional review board approval and informed consent were not required. Two radiologists reviewed CT images from 14 patients (five men, nine women) who had cancer or metastasis whose malignancy classified as GGO. region sampled completely covered with contiguous, 50% overlapping regions interest (ROIs) measuring 30 × pixels size....
PURPOSE: To retrospectively compare performance of artificial neural networks (ANNs) applied to ultrasonographic (US) images with that radiologists for prediction appropriateness a donor liver respect macrosteatosis before transplantation. MATERIALS AND METHODS: Institutional ethics committee approved study; written informed consent was obtained. ANNs, constructed three-layered 15-neuron back-propagation algorithm, were trained predict by using statistically significant laboratory and US...
Purpose: Analyzing spatiotemporal enhancement patterns is an important task for the differential diagnosis of breast tumors in dynamic contrast‐enhanced MRI (DCE‐MRI), and yet remains challenging because complexities analyzing time‐series three‐dimensional image data. The authors propose a novel approach to computer‐aided (CAD) using multilevel analysis association features tumor DCE‐MRI. Methods: A database 171 cases consisting 111 malignant 60 benign was used. Time‐series MR images were...
The purpose of this study was to evaluate the use xenon-enhanced dual-energy CT chest assess ventilation changes after methacholine and salbutamol inhalation in subjects with asthma healthy subjects.Twenty-five 10 underwent three-phase (basal, inhalation, inhalation) CT. Each phase composed wash-in washout scans. For visual analysis, two radiologists evaluated defects gas trapping lobe by on a 10-point scale. Total defect scores were calculated adding scores. Xenon total lung volume...
Objective:The purpose of this study was to assess whether a deep learning (DL) algorithm could enable simultaneous noise reduction and edge sharpening in low-dose lumbar spine CT. Materials Methods: This retrospective included 52 patients (26 male 26 female; median age, 60.5 years) who had undergone CT-guided bone biopsy between October 2015 April 2020.Initial 100-mAs survey images 50-mAs intraprocedural were reconstructed by filtered back projection.Denoising performed using vendor-agnostic...
PURPOSE: To examine the combined effects of monitor luminance and ambient light on observer performance for detecting abnormalities in a soft-copy interpretation digital chest radiographs. MATERIALS AND METHODS: A total 254 radiographs were displayed high-resolution cathode ray tube at three levels (25, 50, 100 foot-lamberts) under (0, 460 lux). Six radiologists reviewed each image nine modes light. The observers allowed to adjust window width level images. included nodule, pneumothorax,...
Capsule Endoscopy(CE) is a new modality for convenient and accurate investigation of small bowel pathology including obscure gastrointestinal bleeding. In this paper, we present method automatic detection the bleeding region high performance CE system with feature index in CE. The transmits 3 images per second to receiver about 10 hours. Over 100,000 are recorded receiver. Typically, viewing such large amount gives clinicians significant burden [3]. We propose technique detect automatically,...
Objective:The purpose of this study was to evaluate the reliability and quality radiomic features in glioblastoma multiforme (GBM) derived from tumor volumes obtained with semi-automated segmentation software.Materials Methods: MR images 45 GBM patients (29 males, 16 females) were downloaded The Cancer Imaging Archive, which post-contrast T1-weighted imaging fluid-attenuated inversion recovery sequences used.Two raters independently segmented tumors using two tools (TumorPrism3D 3D...
Mammographic breast density is a well-established marker for cancer risk. However, accurate measurement of dense tissue difficult task due to faint contrast and significant variations in background fatty tissue. This study presents novel method automated mammographic estimation based on Convolutional Neural Network (CNN). A total 397 full-field digital mammograms were selected from Seoul National University Hospital. Among them, 297 randomly as training set the rest 100 used test set. We...
We sought to evaluate the potential benefits of a computer-aided detection (CAD) system for detecting lung nodules in multidetector row CT (MDCT) scans.A CAD was developed on MDCT scans and applied data obtained from 15 patients. Two chest radiologists consensus established reference standard. The were categorized according their size relationship surrounding structures (nodule type). differences sensitivities between an experienced radiologist without user interaction evaluated using chi2...