• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.06a
• /
• pp.110-110
• /
• 2022

본 연구는 자율운항 선박의 원격 고장 진단 기법 개발의 일부로 수행되었다. 특히, 엔진 연료 계통 장비로부터 계측된 시계열 데이터로부터 상태 진단을 위한 알고리즘 구현 결과를 제시하였다. 엔진 연료 펌프와 청정기를 가진 육상 실험 장비로부터 진동 시계열 데이터 계측하였으며, 이상 감지, 고장 분류 및 고장 예측이 가능한 심층 학습(Deep Learning) 및 기계 학습(Machine Learning) 알고리즘을 구현하였다. 육상 실험 장비에 고장 유형 별로 인위적인 고장을 발생시켜 특징적인 진동 신호를 계측하여, 인공 지능 학습에 이용하였다. 계측된 신호 데이터는 선행 발생한 사건의 신호가 후행 사건에 영향을 미치는 특성을 가지고 있으므로, 시계열에 내포된 고장 상태는 시간 간의 선후 종속성을 반영할 수 있는 학습 알고리즘을 제시하였다. 고장 사건의 시간 종속성을 반영할 수 있도록 순환(Recurrent) 계열의 RNN(Recurrent Neural Networks), LSTM(Long Short-Term Memory models)의 모델과 합성곱 연산 (Convolution Neural Network)을 기반으로 하는 Conv1D 모델을 적용하여 예측 정확성을 비교하였다. 특히, 합성곱 계열의 RNN LSTM 모델이 고차원의 순차적 자연어 언어 처리에 장점을 보이는 모델임을 착안하여, 신호의 시간 종속성을 학습에 반영할 수 있는 합성곱 계열의 Conv1 알고리즘을 고장 예측에 사용하였다. 또한 기계 학습 모델의 효율성을 감안하여 XGBoost를 추가로 적용하여 고장 예측을 시도하였다. 최종적으로 연료 펌프와 청정기의 진동 신호로부터 Conv1D 모델과 XGBoost 모델의 고장 예측 성능 결과를 비교하였다

• Annual Conference of KIPS
• /
• 2023.11a
• /
• 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.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.1055-1065
• /
• 2023

Recently, various studies using deep reinforcement learning (deep RL) technology have been conducted to solve complex problems using big data collected at industrial internet of things. Deep RL uses reinforcement learning"s trial-and-error algorithms and cumulative compensation functions to generate and learn its own data and quickly explore neural network structures and parameter decisions. However, studies so far have shown that the larger the size of the learning data is, the higher are the memory usage and search time, and the lower is the accuracy. In this study, model-agnostic learning for efficient federated deep RL was utilized to solve privacy invasion by increasing robustness as 55.9% and achieve 97.8% accuracy, an improvement of 5.5% compared with the comparative optimization-based meta learning models, and to reduce the delay time by 28.9% on average.

• Imaging Science in Dentistry
• /
• v.53 no.4
• /
• pp.271-281
• /
• 2023

Purpose: The objective of this scoping review was to investigate the applicability and performance of various convolutional neural network (CNN) models in tooth numbering on panoramic radiographs, achieved through classification, detection, and segmentation tasks. Materials and Methods: An online search was performed of the PubMed, Science Direct, and Scopus databases. Based on the selection process, 12 studies were included in this review. Results: Eleven studies utilized a CNN model for detection tasks, 5 for classification tasks, and 3 for segmentation tasks in the context of tooth numbering on panoramic radiographs. Most of these studies revealed high performance of various CNN models in automating tooth numbering. However, several studies also highlighted limitations of CNNs, such as the presence of false positives and false negatives in identifying decayed teeth, teeth with crown prosthetics, teeth adjacent to edentulous areas, dental implants, root remnants, wisdom teeth, and root canal-treated teeth. These limitations can be overcome by ensuring both the quality and quantity of datasets, as well as optimizing the CNN architecture. Conclusion: CNNs have demonstrated high performance in automated tooth numbering on panoramic radiographs. Future development of CNN-based models for this purpose should also consider different stages of dentition, such as the primary and mixed dentition stages, as well as the presence of various tooth conditions. Ultimately, an optimized CNN architecture can serve as the foundation for an automated tooth numbering system and for further artificial intelligence research on panoramic radiographs for a variety of purposes.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.119-126
• /
• 2024

In building information model (BIM), it is difficult to train an artificial intelligence (AI) model due to the lack of sufficient data about individual projects in an architecture firm. In this paper, we present a methodology to correctly train an AI neural network model based on a large language model (LLM) to predict the steel structure product weight ratios in BIM. The proposed method, with the aid of the LLM, can overcome the inherent problem of limited data availability in BIM and handle a combination of natural language and numerical data. The experimental results showed that the proposed method demonstrated significantly higher accuracy than methods based on a smaller language model. The potential for effectively applying large language models in BIM is confirmed, leading to expectations of preventing building accidents and efficiently managing construction costs.

• Computers and Concrete
• /
• v.33 no.4
• /
• pp.341-348
• /
• 2024

Ultra-high-performance concrete (UHPC) has received remarkable attentions in civil infrastructure due to its unique mechanical characteristics and durability. UHPC gains increasingly dominant in essential structural elements, while its unique properties pose challenges for traditional inspection methods, as damage may not always manifest visibly on the surface. As such, the need for robust inspection techniques for detecting cracks in UHPC members has become imperative as traditional methods often fall short in providing comprehensive and timely evaluations. In the era of artificial intelligence, computer vision has gained considerable interest as a powerful tool to enhance infrastructure condition assessment with image and video data collected from sensors, cameras, and unmanned aerial vehicles. This paper presents a computer vision-based approach employing deep learning to detect cracks in UHPC beams, with the aim of addressing the inherent limitations of traditional inspection methods. This work leverages computer vision to discern intricate patterns and anomalies. Particularly, a convolutional neural network architecture employing transfer learning is adopted to identify the presence of cracks in the beams. The proposed approach is evaluated with image data collected from full-scale experiments conducted on UHPC beams subjected to flexural and shear loadings. The results of this study indicate the applicability of computer vision and deep learning as intelligent methods to detect major and minor cracks and recognize various damage mechanisms in UHPC members with better efficiency compared to conventional monitoring methods. Findings from this work pave the way for the development of autonomous infrastructure health monitoring and condition assessment, ensuring early detection in response to evolving structural challenges. By leveraging computer vision, this paper contributes to usher in a new era of effectiveness in autonomous crack detection, enhancing the resilience and sustainability of UHPC civil infrastructure.

• Steel and Composite Structures
• /
• v.51 no.4
• /
• pp.441-456
• /
• 2024

This paper aims to develop Machine Learning (ML) algorithms to predict the buckling resistance of cold-formed steel (CFS) channels with restrained flanges, widely used in typical CFS sheathed wall panels, and provide practical design tools for engineers. The effects of cross-sectional restraints were first evaluated on the elastic buckling behaviour of CFS channels subjected to pure axial compressive load or bending moment. Feedforward multi-layer Artificial Neural Networks (ANNs) were then trained on different datasets comprising CFS channels with various dimensions and properties, plate thicknesses, and restraining conditions on one or two flanges, while the elastic distortional buckling resistance of the elements were determined according to the Finite Strip Method (FSM). To develop less biased networks and ensure that every observation from the original dataset has the chance of appearing in the training and test set, a K-fold cross-validation technique was implemented. In addition, the hyperparameters of the ANNs were tuned using a grid search technique to provide ANNs with optimum performances. The results demonstrated that the trained ANNs were able to predict the elastic distortional buckling resistance of CFS flange-restrained elements with an average accuracy of 99% in terms of coefficient of determination. The developed models were then used to propose a simple ANN-based design formula for the prediction of the elastic distortional buckling stress of CFS flange-restrained elements. Finally, the proposed formula was further evaluated on a separate set of unseen data to ensure its accuracy for practical applications.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.3
• /
• pp.109-123
• /
• 2024

Research into vision-based end-to-end autonomous driving systems utilizing deep learning and reinforcement learning has been steadily increasing. These systems typically encode continuous and high-dimensional vehicle states, such as location, velocity, orientation, and sensor data, into latent features, which are then decoded into a vehicular control policy. The complexity of urban driving environments necessitates the use of state representation learning through networks like Variational Autoencoders (VAEs) or Convolutional Neural Networks (CNNs). This paper analyzes the impact of different image state encoding methods on reinforcement learning performance in autonomous driving. Experiments were conducted in the CARLA simulator using RGB images and semantically segmented images captured by the vehicle's front camera. These images were encoded using VAE and Vision Transformer (ViT) networks. The study examines how these networks influence the agents' learning outcomes and experimentally demonstrates the role of each state representation technique in enhancing the learning efficiency and decision- making capabilities of autonomous driving systems.

• Structural Engineering and Mechanics
• /
• v.91 no.6
• /
• pp.643-656
• /
• 2024

The prediction of the moment rotation behavior of semi-rigid connections has been the subject of extensive research. However, to improve the accuracy of these predictions, there is a growing interest in employing machine learning algorithms. This paper investigates the effectiveness of using Multi-gene genetic programming (MGGP) to predict the moment-rotation behavior of flush-end plate connections compared to that of artificial neural networks (ANN) and previous studies. It aims to automate the process of determining the most suitable equations to accurately describe the behavior of these types of connections. Experimental data was used to train ANN and MGGP. The performance of the models was assessed by comparing the values of coefficient of determination (R²), maximum absolute error (MAE), and root-mean-square error (RMSE). The results showed that MGGP produced more accurate, reliable, and general predictions compared to ANN and previous studies with an R² exceeding 0.99, an RMSE of 6.97, and an MAE of 38.68, highlighting its advantages over other models. The use of MGGP can lead to better modeling and more precise predictions in structural design. Additionally, an experimentally-based regression analysis was conducted to obtain the rotational capacity of FECs. A new equation was proposed and compared to previous ones, showing significant improvement in accuracy with an R² score of 0.738, an RMSE of 0.014, and an MAE of 0.024.

• Nuclear Engineering and Technology
• /
• v.56 no.8
• /
• pp.3123-3128
• /
• 2024

The dicentric chromosome assay is a gold standard method to estimate radiation exposure by calculating the ratio of dicentric chromosomes existing in cells. The objective of this study was to propose an automatic dicentric chromosome discrimination method based on deep convolutional neural networks using radiation exposure patient data. From 45 patients with radiation exposure, conventional Giemsa-stained images of 116,258 normal and 2800 dicentric chromosomes were confirmed. ImageNet was used to pre-train VGG19, which was modified and fine-tuned. The proposed modified VGG19 demonstrated dicentric chromosome discrimination performance, with a true positive rate of 0.927, a true negative rate of 0.997, a positive predictive value of 0.882, a negative predictive value of 0.998, and an area under the receiver operating characteristic curve of 0.997.