• Korean Journal of Radiology
• /
• 제24권11호
• /
• pp.1061-1080
• /
• 2023

Artificial intelligence (AI) in radiology is a rapidly developing field with several prospective clinical studies demonstrating its benefits in clinical practice. In 2022, the Korean Society of Radiology held a forum to discuss the challenges and drawbacks in AI development and implementation. Various barriers hinder the successful application and widespread adoption of AI in radiology, such as limited annotated data, data privacy and security, data heterogeneity, imbalanced data, model interpretability, overfitting, and integration with clinical workflows. In this review, some of the various possible solutions to these challenges are presented and discussed; these include training with longitudinal and multimodal datasets, dense training with multitask learning and multimodal learning, self-supervised contrastive learning, various image modifications and syntheses using generative models, explainable AI, causal learning, federated learning with large data models, and digital twins.

• 폴리머
• /
• 제38권6호
• /
• pp.815-819
• /
• 2014

조직공학에서의 인공지지체는 세포의 부착과 증식 및 분화가 잘 되어야 하고, 우수한 생체친화성 및 생분해성을 지녀야 한다. 다양한 인공지지체 제작 방법이 시도되어지고 있으며, 최근들어 3D 프린팅 기술을 이용한 방식이 활발하게 연구되어지고 있다. 폴리카프로락톤(polycaprolactone, PCL)은 낮은 녹는점을 가지고 있어 3D 프린팅하기에 우수한 생체적합 고분자 합성재료이다. 본 연구에서는 3D 프린팅 기술을 이용하여 3차원 PCL 인공지지체를 제작하였고, 지지체의 표면개질을 위해 수산화나트륨(NaOH)을 이용하였다. 표면개질된 인공지지체의 표면특성을 SEM으로 확인한 결과, 수산화나트륨을 처리한 PCL 인공지지체가 처리하지 않은 PCL 인공지지체에 비해 거칠기가 증가함을 보였으며, 접촉각 측정을 통해 친수성이 증가함을 확인하였다. In vitro 실험결과, 수산화나트륨을 처리한 PCL 인공지지체가 처리하지 않은 PCL 인공지지체에 비해 세포의 증식과 분화가 증가함을 보였고, 세포의 부착 모습은 균일하고 밀집된 형태로 부착됨을 확인하였다. 따라서 조형가공기술을 이용하여 수산화나트륨을 처리한 표면개질된 PCL 인공지지체를 제작하고 분석함으로써, 세포적합성을 통해 체내 인공지지체 개발 적용 가능성을 제시하였다.

• 대한의용생체공학회:의공학회지
• /
• 제44권3호
• /
• pp.211-217
• /
• 2023

Recently, many studies have introduced artificial intelligence systems in the surgical process to reduce the incidence and mortality of complications in patients. Bleeding is a major cause of operative mortality and complications. However, there have been few studies conducted on detecting bleeding in surgical videos. To advance the development of deep learning models for detecting intraoperative hemorrhage, three models have been trained and compared; such as, YOLOv5, RetinaNet50, and RetinaNet101. We collected 1,016 bleeding images extracted from five surgical videos. The ground truths were labeled based on agreement from two specialists. To train and evaluate models, we divided the datasets into training data, validation data, and test data. For training, 812 images (80%) were selected from the dataset. Another 102 images (10%) were used for evaluation and the remaining 102 images (10%) were used as the evaluation data. The three main metrics used to evaluate performance are precision, recall, and false positive per image (FPPI). Based on the evaluation metrics, RetinaNet101 achieved the best detection results out of the three models (Precision rate of 0.99±0.01, Recall rate of 0.93±0.02, and FPPI of 0.01±0.01). The information on the bleeding detected in surgical videos can be quickly transmitted to the operating room, improving patient outcomes.

• 한국융합학회논문지
• /
• 제11권8호
• /
• pp.1-6
• /
• 2020

수박의 맛을 평가하는 다양한 방식이 있으나, 기존의 방법들은 주관적 방식, 평가 비용, 대상의 손상 등과 같은 평가 방식의 한계점이 있다. 최근에는 이러한 단점들을 해소하기 위해 소리를 이용하여 수박을 평가하는 연구들이 진행되고 있다. 본 연구에서는 수박을 두드렸을 때 나는 반향 소리를 AI기반의 기계 학습을 이용하여 수박의 당도를 예측하는 모델을 개발 하였다. 수박의 당도가 높을수록 높은 주파수 성분이 특이점으로 나타나며, 따라서 반향소리 시간-주파수 특이점에 기반 하여 기계 학습 방법을 개발하였다. 2개의 수박 당도별 그룹을 구분 시에 83.2%, 3개의 그룹을 구분시에 59.6%의 정확도로 당도를 예측 할 수 있었다.

• 한국분말재료학회지
• /
• 제24권2호
• /
• pp.96-101
• /
• 2017

Well-dispersed platinum catalysts on ruthenium oxide nanofiber supports are fabricated using electrospinning, post-calcination, and reduction methods. To obtain the well-dispersed platinum catalysts, the surface of the nanofiber supports is modified using post-calcination. The structures, morphologies, crystal structures, chemical bonding energies, and electrochemical performance of the catalysts are investigated. The optimized catalysts show well-dispersed platinum nanoparticles (1-2 nm) on the nanofiber supports as well as a uniform network structure. In particular, the well-dispersed platinum catalysts on the ruthenium oxide nanofiber supports display excellent catalytic activity for oxygen reduction reactions with a half-wave potential ($E_{1/2}$) of 0.57 V and outstanding long-term stability after 2000 cycles, resulting in a lower $E_{1/2}$ potential degradation of 19 mV. The enhanced electrochemical performance for oxygen reduction reactions results from the well-dispersed platinum catalysts and unique nanofiber supports.

• 한국분말재료학회지
• /
• 제26권1호
• /
• pp.58-69
• /
• 2019

High performance lithium-ion batteries (LIBs) have attracted considerable attention as essential energy sources for high-technology electrical devices such as electrical vehicles, unmanned drones, uninterruptible power supply, and artificial intelligence robots because of their high energy density (150-250 Wh/kg), long lifetime (> 500 cycles), low toxicity, and low memory effects. Of the high-performance LIB components, cathode materials have a significant effect on the capacity, lifetime, energy density, power density, and operating conditions of high-performance LIBs. This is because cathode materials have limitations with respect to a lower specific capacity and cycling stability as compared to anode materials. In addition, cathode materials present difficulties when used with LIBs in electric vehicles because of their poor rate performance. Therefore, this study summarizes the structural and electrochemical properties of cathode materials for LIBs used in electric vehicles. In addition, we consider unique strategies to improve their structural and electrochemical properties.

• Journal of Electrochemical Science and Technology
• /
• 제6권3호
• /
• pp.106-110
• /
• 2015

This study investigates Cu(II) complexes of cyclam, propylene cross-bridged cyclam (PCB-cyclam), and propylene cross-bridged cyclam diacetate (PCB-TE2A) as homogeneous electrocatalysts for CO₂ reduction in comparison with Ni(II)-cyclam. It is found that Cu(II)-cyclam can catalyze CO₂ reduction at the potential close to its thermodynamic value (0.75 V vs. Ag/AgCl) in tris-HCl buffer (pH 8.45) on a glassy carbon electrode. Cu(II)-cyclam, however, suffers from severe demetalation due to the insufficient stability of Cu(I)-cyclam. Cu(II)-PCB-cyclam and Cu(II)-PCB-TE2A are revealed to exhibit much less demetalation behavior, but poor CO₂ reduction activities as well. The inferior electrocatalytic ability of Cu(II)-PCB-cyclam is ascribed to its redox potential that is too high for CO₂ reduction, and that of Cu(II)-PCB-TE2A to the steric hindrance preventing facile contact with CO₂ molecules. This study suggests that in addition to the redox potential and chemical stability, the stereochemical aspect has to be considered in designing efficient electrocatalysts for CO₂ reduction.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2023년도 추계학술발표대회
• /
• pp.685-686
• /
• 2023

상수도관의 수질 및 부식도 검사에는 파이프에 손상을 입히지 않고 지속적인 방법이 필요하다. 초음파는 이를 만족하면서 상태를 확인할 수 있고 주파수가 높을수록 해상도가 좋아져 정밀한 측정이 가능하다는 장점이 있다. 이러한 특성을 이용해 상수도관 모니터링 시스템으로 초음파 기반의 Scanning Acoustic Microscopy(SAM)과 Convolutional Neural Network(CNN)을 사용하는 새로운 방법을 제안한다. 기존의 Non-Destructive Testing(NDT)방식의 단점을 보완하면서 더 높은 해상도로 상수도관을 점검하는 방식으로, SAM 을 이용하여 부식으로 인한 파이프 두께 변화와 부유물의 여부 및 수질을 동시에 감지하고 얻은 데이터를 CNN 으로 분석했다. CNN 의 높은 정확도 결과로 이 시스템의 파이프 부식도 및 수질 모니터링에 대한 적합성을 보여주었다.

• International journal of advanced smart convergence
• /
• 제13권2호
• /
• pp.195-204
• /
• 2024

In this paper, we investigate the comparative performance of two leading object detection architectures, YOLOv7 and Detection Transformer (DETR), across varying levels of data augmentation using CycleGAN. Our experiments focus on chest scan images within the context of biomedical informatics, specifically targeting the detection of abnormalities. The study reveals that YOLOv7 consistently outperforms DETR across all levels of augmented data, maintaining better performance even with 75% augmented data. Additionally, YOLOv7 demonstrates significantly faster convergence, requiring approximately 30 epochs compared to DETR's 300 epochs. These findings underscore the superiority of YOLOv7 for object detection tasks, especially in scenarios with limited data and when rapid convergence is essential. Our results provide valuable insights for researchers and practitioners in the field of computer vision, highlighting the effectiveness of YOLOv7 and the importance of data augmentation in improving model performance and efficiency.

• Molecules and Cells
• /
• 제37권12호
• /
• pp.888-898
• /
• 2014

Pancreatic cancer is one of the most fatal cancers and is associated with limited diagnostic and therapeutic modalities. Currently, gemcitabine is the only effective drug and represents the preferred first-line treatment for chemotherapy. However, a high level of intrinsic or acquired resistance of pancreatic cancer to gemcitabine can contribute to the failure of gemcitabine treatment. To investigate the underlying molecular mechanisms for gemcitabine resistance in pancreatic cancer, we performed label-free quantification of protein expression in intrinsic gemcitabine-resistant and -sensitive human pancreatic adenocarcinoma cell lines using our improved proteomic strategy, combined with filter-aided sample preparation, single-shot liquid chromatography-mass spectrometry, enhanced spectral counting, and a statistical method based on a power law global error model. We identified 1931 proteins and quantified 787 differentially expressed proteins in the BxPC3, PANC-1, and HPDE cell lines. Bioinformatics analysis identified 15 epithelial to mesenchymal transition (EMT) markers and 13 EMT-related proteins that were closely associated with drug resistance were differentially expressed. Interestingly, 8 of these proteins were involved in glutathione and cysteine/methionine metabolism. These results suggest that proteins related to the EMT and glutathione metabolism play important roles in the development of intrinsic gemcitabine resistance by pancreatic cancer cell lines.