• 한국전자통신학회논문지
• /
• 제19권1호
• /
• pp.267-274
• /
• 2024

본 논문에서는 스마트공장에 관한 연도별 국내 특허 출원 건수, 연도별 국내 특허 공개 건수 및 연도별 국내 등록 건수를 조사하였고, 출원인 유형별 특허 출원 건수를 조사하였다. 조사된 특허를 기반으로 가장 많은 특허에서 나타나는 IPC는 G05B 19/418이라는 것을 밝혀냈다. 또한 스마트공장 특허 IPC의 사회 연결망 분석을 통해 G05B 19/418이 연결중심성이 가장 높은 IPC임을 밝혀냈다. 이상의 내용을 통해 스마트공장으로 제출된 특허의 핵심기술의 IPC가 G05B 19/418일 경우, G05B 23/02와 조합된 기술 즉, "공장 제어"와 "모니터링"이 복합된 기술이 특허로 가장 많이 출원되었다는 것과 핵심기술의 IPC가 G06Q 50/04일 경우, G06Q 50/10과 조합된 기술 즉, "제조"와 "서비스"가 복합된 기술이 특허로 가장 많이 출원되었다는 것을 확인할 수 있었다. 이를 통해 스마트공장으로 특허를 출원하기 위해서는 이러한 IPC간의 연결성 고려한 특허출원이 필요한 것을 알 수 있었다.

• Korean Journal of Radiology
• /
• 제24권4호
• /
• pp.294-304
• /
• 2023

Objective: We aimed to investigate whether image standardization using deep learning-based computed tomography (CT) image conversion would improve the performance of deep learning-based automated hepatic segmentation across various reconstruction methods. Materials and Methods: We collected contrast-enhanced dual-energy CT of the abdomen that was obtained using various reconstruction methods, including filtered back projection, iterative reconstruction, optimum contrast, and monoenergetic images with 40, 60, and 80 keV. A deep learning based image conversion algorithm was developed to standardize the CT images using 142 CT examinations (128 for training and 14 for tuning). A separate set of 43 CT examinations from 42 patients (mean age, 10.1 years) was used as the test data. A commercial software program (MEDIP PRO v2.0.0.0, MEDICALIP Co. Ltd.) based on 2D U-NET was used to create liver segmentation masks with liver volume. The original 80 keV images were used as the ground truth. We used the paired t-test to compare the segmentation performance in the Dice similarity coefficient (DSC) and difference ratio of the liver volume relative to the ground truth volume before and after image standardization. The concordance correlation coefficient (CCC) was used to assess the agreement between the segmented liver volume and ground-truth volume. Results: The original CT images showed variable and poor segmentation performances. The standardized images achieved significantly higher DSCs for liver segmentation than the original images (DSC [original, 5.40%-91.27%] vs. [standardized, 93.16%-96.74%], all P < 0.001). The difference ratio of liver volume also decreased significantly after image conversion (original, 9.84%-91.37% vs. standardized, 1.99%-4.41%). In all protocols, CCCs improved after image conversion (original, -0.006-0.964 vs. standardized, 0.990-0.998). Conclusion: Deep learning-based CT image standardization can improve the performance of automated hepatic segmentation using CT images reconstructed using various methods. Deep learning-based CT image conversion may have the potential to improve the generalizability of the segmentation network.

• Korean Journal of Radiology
• /
• 제23권10호
• /
• pp.949-958
• /
• 2022

Objective: To investigate the feasibility of using a deep learning-based analysis of auscultation data to predict significant stenosis of arteriovenous fistulas (AVF) in patients undergoing hemodialysis requiring percutaneous transluminal angioplasty (PTA). Materials and Methods: Forty patients (24 male and 16 female; median age, 62.5 years) with dysfunctional native AVF were prospectively recruited. Digital sounds from the AVF shunt were recorded using a wireless electronic stethoscope before (pre-PTA) and after PTA (post-PTA), and the audio files were subsequently converted to mel spectrograms, which were used to construct various deep convolutional neural network (DCNN) models (DenseNet201, EfficientNetB5, and ResNet50). The performance of these models for diagnosing ≥ 50% AVF stenosis was assessed and compared. The ground truth for the presence of ≥ 50% AVF stenosis was obtained using digital subtraction angiography. Gradient-weighted class activation mapping (Grad-CAM) was used to produce visual explanations for DCNN model decisions. Results: Eighty audio files were obtained from the 40 recruited patients and pooled for the study. Mel spectrograms of "pre-PTA" shunt sounds showed patterns corresponding to abnormal high-pitched bruits with systolic accentuation observed in patients with stenotic AVF. The ResNet50 and EfficientNetB5 models yielded an area under the receiver operating characteristic curve of 0.99 and 0.98, respectively, at optimized epochs for predicting ≥ 50% AVF stenosis. However, Grad-CAM heatmaps revealed that only ResNet50 highlighted areas relevant to AVF stenosis in the mel spectrogram. Conclusion: Mel spectrogram-based DCNN models, particularly ResNet50, successfully predicted the presence of significant AVF stenosis requiring PTA in this feasibility study and may potentially be used in AVF surveillance.

• 한국전산구조공학회논문집
• /
• 제37권2호
• /
• pp.77-84
• /
• 2024

4차 산업혁명 시대에 건설산업은 전통적인 업무 방식에서 디지털 프로세스로 전환하고 있다. 특히, 건설산업의 특성으로 인해 업무 절차의 변경에는 어려움이 따르며, 점진적인 디지털 전환 및 시행착오가 발생하고 있다. 건설현장의 안전관리 분야도 역시 이 흐름을 따라 모든 데이터의 디지털화와 자동화를 목표로 연구 및 시도가 활발히 진행되고 있다. 그러나 최근의 통계에 따르면, 건설업 안전사고는 계속해서 발생하고 있으며, 안전사고 사망자 수도 줄지 않고 있다. 본 연구는 이러한 문제를 해결하기 위해 건설공사 안전관리 종합정보망의 빅데이터를 대규모 언어모델 인공지능을 통해 분석하였다. 분석된 결과는 실시간으로 업데이트가 가능한 상세설계모델로부터 위치정보와 공간적 특성을 반영하여 안전관리가 필요한 현장모델링에 정보를 맵핑하였다. 해당 연구를 통해 건설현장 안전관리 데이터의 디지털화를 통한 시설물 및 근로자의 안전을 강화하고, 건설사고 예방 및 효과적인 교육 지시를 위한 빅데이터 기반 안전관리 플랫폼 개발을 목표로 한다.

• 한국수자원학회논문집
• /
• 제57권4호
• /
• pp.277-287
• /
• 2024

상수도 관망은 운영 중 다양한 수질 사고 발생 위험요소에 노출되어 있다. 본 연구는 다양한 수질 사고 위험요소 중 상수도 관망 내로의 바이러스 유입에 따른 위험도 평가 방법론을 제시하고, 이를 활용하여 위험도를 최소화할 수 있는 재염소 시설의 위치와 운영에 대한 검토를 수행하였다. 위험도 평가를 위해 QMRA (Quantitative Microbial Risk assessment)를 적용하였으며, 염소 농도에 따른 Water Quality Resilience를 정의하여 바이러스가 유입되지 않은 정상 운영 상황과 바이러스가 유입된 비정상 상황에서 염소 농도가 목표 범위(0.1-1.0 mg/L)내 운영되는지 여부를 정량적으로 확인하였다. 본 연구에서는 바이러스와 염소간의 반응을 고려해야 하기에, 다양한 수질인자를 고려할 수 있는 EPANET-MSX를 활용하여 수리 및 수질 분석을 수행하였다. 제안한 방법론은 미국의 Bellingham의 관망에 적용하였으며, 재염소 시설의 경우 0.5 mg/L부터 1.0 mg/L까지 주입 가능한 것으로 하였다. 적용 결과 재염소 시설이 없는 경우 Average risk가 0.0154이었으며, 재염소 시설 설치 후 1.0 mg/L의 농도로 주입 시 39.1%의 Average risk를 저감할 수 있었다. 다만, 재염소 시설을 통한 과도한 염소 주입은 Water Quality Resilience를 저감하여, 최종적으로 0.5 mg/L의 재염소 시설을 선정하였으며, 이를 활용하여 20% 가량의 Average risk 저감이 가능함을 확인하였다. 본 연구는 향후 잠재적 바이러스 유입에 대비한 재염소 시설의 설계에 활용할 수 있을 것이다.

• 한국컴퓨터정보학회논문지
• /
• 제29권2호
• /
• pp.31-41
• /
• 2024

딥러닝 모델(Deep Learning Model)은 컴퓨터 비전(Computer Vision) 분야의 이미지(Image) 분류 및 객체 탐지와 같은 작업에서 뛰어난 성과를 보이며, 실제 산업 현장에서 다양하게 활용되고 있다. 최근 다양한 알고리즘(Algorithm)의 적대적 예제를 이용하여 딥러닝 모델의 취약성을 지적하며, 강건성 향상 방안을 제시하는 연구들이 활발하게 진행되고 있다. 적대적 예제는 오분류를 유도하기 위해 작은 노이즈(Noise)가 추가된 이미지로서, 딥러닝 모델을 실제 환경에 적용 시 중대한 위협이 될 수 있다. 본 논문에서는 다양한 알고리즘의 적대적 예제를 대상으로 에지 학습 분류 모델의 강건성 및 이를 이용한 적대적 예제 탐지 모델의 성능을 확인하고자 하였다. 강건성 실험 결과, FGSM(Fast Gradient Sign Method) 알고리즘에 대하여 기본 분류 모델이 약 17%의 정확도를 보였으나, 에지(Edge) 학습 모델들은 60~70%대의 정확도를 유지하였고, PGD(projected gradient descent)/DeepFool/CW(Carlini-Wagner) 알고리즘에 대해서는 기본 분류 모델이 0~1%의 정확도를 보였으나, 에지 학습 모델들은 80~90%의 정확도를 유지하였다. 적대적 예제 탐지 실험 결과, FGSM/PGD/DeepFool/CW의 모든 알고리즘에 대해서 91~95%의 높은 탐지율을 확인할 수 있었다. 본 연구를 통하여 다양한 적대적 알고리즘에 대한 방어 가능성을 제시함으로써, 컴퓨터 비전을 활용하는 여러 산업 분야에서 딥러닝 모델의 안전성 및 신뢰성 제고를 기대한다.

• IMMUNE NETWORK
• /
• 제23권5호
• /
• pp.39.1-39.15
• /
• 2023

Coronavirus disease 2019 (COVID-19) vaccination may non-specifically alter the host immune system. This study aimed to evaluate the effect of COVID-19 vaccination on hepatitis B surface Ag (HBsAg) titer and host immunity in chronic hepatitis B (CHB) patients. Consecutive 2,797 CHB patients who had serial HBsAg measurements during antiviral treatment were included in this study. Changes in the HBsAg levels after COVID-19 vaccination were analyzed. The dynamics of NK cells following COVID-19 vaccination were also examined using serial blood samples collected prospectively from 25 healthy volunteers. Vaccinated CHB patients (n=2,329) had significantly lower HBsAg levels 1-30 days post-vaccination compared to baseline (median, -21.4 IU/ml from baseline), but the levels reverted to baseline by 91-180 days (median, -3.8 IU/ml). The velocity of the HBsAg decline was transiently accelerated within 30 days after vaccination (median velocity: -0.06, -0.39, and -0.04 log₁₀ IU/ml/year in pre-vaccination period, days 1-30, and days 31-90, respectively). In contrast, unvaccinated patients (n=468) had no change in HBsAg levels. Flow cytometric analysis showed that the frequency of NK cells expressing NKG2A, an NK inhibitory receptor, significantly decreased within 7 days after the first dose of COVID-19 vaccine (median, -13.1% from baseline; p<0.001). The decrease in the frequency of NKG2A⁺ NK cells was observed in the CD56^dimCD16⁺ NK cell population regardless of type of COVID-19 vaccine. COVID-19 vaccination leads to a rapid, transient decline in HBsAg titer and a decrease in the frequency of NKG2A⁺ NK cells.

• IMMUNE NETWORK
• /
• 제23권6호
• /
• pp.43.1-43.10
• /
• 2023

The continuous emergence of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) variants has provided insights for updating current coronavirus disease 2019 (COVID-19) vaccines. We examined the neutralizing activity of Abs induced by a BA.4/5-containing bivalent mRNA vaccine against Omicron subvariants BN.1 and XBB.1.5. We recruited 40 individuals who had received a monovalent COVID-19 booster dose after a primary series of COVID-19 vaccinations and will be vaccinated with a BA.4/5-containing bivalent vaccine. Sera were collected before vaccination, one month after, and three months after a bivalent booster. Neutralizing Ab (nAb) titers were measured against ancestral SARS-CoV-2 and Omicron subvariants BA.5, BN.1, and XBB.1.5. BA.4/5-containing bivalent vaccination significantly boosted nAb levels against both ancestral SARS-CoV-2 and Omicron subvariants. Participants with a history of SARS-CoV-2 infection had higher nAb titers against all examined strains than the infection-naïve group. NAb titers against BN.1 and XBB.1.5 were lower than those against the ancestral SARS-CoV-2 and BA.5 strains. These results suggest that COVID-19 vaccinations specifically targeting emerging Omicron subvariants, such as XBB.1.5, may be required to ensure better protection against SARS-CoV-2 infection, especially in high-risk groups.

• IMMUNE NETWORK
• /
• 제23권4호
• /
• pp.35.1-35.16
• /
• 2023

Defining the molecular dynamics associated with T cell differentiation enhances our understanding of T cell biology and opens up new possibilities for clinical implications. In this study, we investigated the dynamics of CD5 expression in CD8⁺ T cell differentiation and explored its potential clinical uses. Using PBMCs from 29 healthy donors, we observed a stepwise decrease in CD5 expression as CD8⁺ T cells progressed through the differentiation stages. Interestingly, we found that CD5 expression was initially upregulated in response to T cell receptor stimulation, but diminished as the cells underwent proliferation, potentially explaining the differentiation-associated CD5 downregulation. Based on the proliferation-dependent downregulation of CD5, we hypothesized that relative CD5 expression could serve as a marker to distinguish the heterogeneous CD8⁺ T cell population based on their proliferation history. In support of this, we demonstrated that effector memory CD8⁺ T cells with higher CD5 expression exhibited phenotypic and functional characteristics resembling less differentiated cells compared to those with lower CD5 expression. Furthermore, in the retrospective analysis of PBMCs from 30 non-small cell lung cancer patients, we found that patients with higher CD5 expression in effector memory T cells displayed CD8⁺ T cells with a phenotype closer to the less differentiated cells, leading to favorable clinical outcomes in response to immune checkpoint inhibitor (ICI) therapy. These findings highlight the dynamics of CD5 expression as an indicator of CD8⁺ T cell differentiation status, and have implications for the development of predictive biomarker for ICI therapy.

• IMMUNE NETWORK
• /
• 제23권6호
• /
• pp.45.1-45.22
• /
• 2023

Interstitial lung disease (ILD) involves persistent inflammation and fibrosis, leading to respiratory failure and even death. Adult tissue-derived mesenchymal stem cells (MSCs) show potential in ILD therapeutics but obtaining an adequate quantity of cells for drug application is difficult. Daewoong Pharmaceutical's MSCs (DW-MSCs) derived from embryonic stem cells sustain a high proliferative capacity following long-term culture and expansion. The aim of this study was to investigate the therapeutic potential of DW-MSCs in experimental mouse models of ILD. DW-MSCs were expanded up to 12 passages for in vivo application in bleomycin-induced pulmonary fibrosis and collagen-induced connective tissue disease-ILD mouse models. We assessed lung inflammation and fibrosis, lung tissue immune cells, fibrosis-related gene/protein expression, apoptosis and mitochondrial function of alveolar epithelial cells, and mitochondrial transfer ability. Intravenous administration of DWMSCs consistently improved lung fibrosis and reduced inflammatory and fibrotic markers expression in both models across various disease stages. The therapeutic effect of DW-MSCs was comparable to that following daily oral administration of nintedanib or pirfenidone. Mechanistically, DW-MSCs exhibited immunomodulatory effects by reducing the number of B cells during the early phase and increasing the ratio of Tregs to Th17 cells during the late phase of bleomycin-induced pulmonary fibrosis. Furthermore, DW-MSCs exhibited anti-apoptotic effects, increased cell viability, and improved mitochondrial respiration in alveolar epithelial cells by transferring their mitochondria to alveolar epithelial cells. Our findings indicate the strong potential of DW-MSCs in the treatment of ILD owing to their high efficacy and immunomodulatory and anti-apoptotic effects.