• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.3
• /
• pp.119-127
• /
• 2019

In this study, we investigated the application of deep learning to the manufacturing process of traffic and road signs which are constituting the road layer in map production with 1 / 1,000 digital topographic map. Automated classification of road traffic sign images was carried out through construction of training data for images acquired by using transfer learning which is used in image classification of deep learning. As a result of the analysis, the signs of attention, regulation, direction and assistance were irregular due to various factors such as the quality of the photographed images and sign shape, but in the case of the guide sign, the accuracy was higher than 97%. In the digital mapping, it is expected that the automatic image classification method using transfer learning will increase the utilization in data acquisition and classification of various layers including traffic safety signs.

• Journal of Broadcast Engineering
• /
• v.25 no.6
• /
• pp.898-910
• /
• 2020

Single image-based super-resolution has been studied for a long time in computer vision because of various applications. Various deep learning-based super-resolution algorithms are introduced recently to improve the performance by reducing side effects like blurring and staircase effects. Most deep learning-based approaches have focused on how to implement the network architecture, loss function, and training strategy to improve performance. Meanwhile, Several approaches using Attention Module, which emphasizes the extracted features, are introduced to enhance the performance of the network without any additional layer. Attention module emphasizes or scales the feature map for the purpose of the network from various perspectives. In this paper, we propose the various channel attention and spatial attention in single image-based super-resolution and analyze the results and performance according to the architecture of the attention module. Also, we explore that designing multi-attention module to emphasize features efficiently from various perspectives.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.27 no.4
• /
• pp.521-528
• /
• 2021

In this study, to repair damaged economizer fin tubes on ships, sealing treatment was performed after applying arc thermal spray coating technology using Inconel 625. A solid particle erosion (SPE) experiment was conducted according to ASTM G76-05 to evaluate the durability of the substrate, thermal spray coating (TSC), and thermal spray coating+sealing treatment (TSC+Sealing) specimens. The surface damage shape was observed using a scanning electron microscope and 3D laser microscope, and the durability was evaluated through the weight loss and surface roughness analysis. Consequently, the durability of the substrate was superior to that of TSC and TSC+Sealing, which was believed to be owing to numerous pore defects in the TSC layer. In addition, the mechanism of solid particle erosion damage was accompanied by plastic deformation and fatigue, which were the characteristics of ductile materials in the case of the substrate, and the tendency of brittle fracture in the case of TSC and TSC+Sealing was confirmed.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.27 no.1
• /
• pp.187-192
• /
• 2021

In this study, Inconel 625 was used as a thermal spray material to prevent dew point corrosion damage to the economizer tube, and sealing treatment was performed after applying the arc thermal spray coating technology. Various electrochemical experiments were conducted in the 0.5 wt% sulfuric acid solution to analyze the corrosion resistance of the thermal spray coating (TSC) layer. After the anodic polarization experiment, the degree of corrosion damage was determined through a scanning electron microscope and EDS component analysis. When measuring the open circuit potential, the effect of the sealing treatment was confirmed through stable potential formation of the TSC+sealing treatment (TSC+Sealing). As a result of the anodic polarization experiment, the passivation region was confirmed in TSC and TSC+Sealing, and corrosion resistance was improved as no corrosion damage was observed. In addition, the corrosion resistance of TSC+Sealing was the best when analyzing the corrosion potential and corrosion current density calculated by Tafel analysis.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• v.25 no.3
• /
• pp.83-95
• /
• 2022

This study aims to present primary data for habitat restoration and artificial breeding conditions of L. unmunsana by identifying the habitat conditions and the larvae's food sources. In order to investigate the habitat characteristics of the adult L. unmunsana and land snails, which are the primary food sources for the larvae, field surveys were conducted on a total of 10 habitats in south-central parts of Korea including Sanseongcheon, Jeonju. The results revealed that the L. unmunsana habitat in the Sanseongcheon area had a broadleaf forest with a multi-layered vegetation structure, adjacent water features, and the north/northeast/northwest slopes with little effect of artificial lighting. The adult L. unmunsana in the Sanseongcheon area appeared from the end of May to the end of June, and was especially intensively observed around the middle of June. The most active time was from 23:30 to 00:30 with a temperature range of 19~22℃ and higher than 80% humidity. The peak count of the observed adults L. unmunsana was a total of 774 on June 11, 2021. In the case of land snails, 11 families and 23 species were observed in 10 habitats of L. unmunsana, and Euphaedusa fusaniana was the most extensive and the most observed in the five survey areas. The land snails of L. unmunsana habitats are mostly found under the organic layers of leaves and a fallen tree branch in broadleaf forests, where a thick organic material layer buffers temperature changes and provides high humidity for various snails. These habitat conditions are suitable for the larva of L. unmunsana and land snails to inhabit, feed, hide and hibernate.

• Structural Engineering and Mechanics
• /
• v.83 no.3
• /
• pp.327-340
• /
• 2022

The Collapse Margin Ratio (CMR) is a notable index used for seismic assessment of the structures. As proposed by FEMA P695, a set of analyses including the Nonlinear Static Analysis (NSA), Incremental Dynamic Analysis (IDA), together with Fragility Analysis, which are typically time-taking and computationally unaffordable, need to be conducted, so that the CMR could be obtained. To address this issue and to achieve a quick and efficient method to estimate the CMR, the Artificial Neural Network (ANN), Response Surface Method (RSM), and Adaptive Neuro-Fuzzy Inference System (ANFIS) will be introduced in the current research. Accordingly, using the NSA results, an attempt was made to find a fast and efficient approach to derive the CMR. To this end, 5016 IDA analyses based on FEMA P695 methodology on 114 various Reinforced Concrete (RC) frames with 1 to 12 stories have been carried out. In this respect, five parameters have been used as the independent and desired inputs of the systems. On the other hand, the CMR is regarded as the output of the systems. Accordingly, a double hidden layer neural network with Levenberg-Marquardt training and learning algorithm was taken into account. Moreover, in the RSM approach, the quadratic system incorporating 20 parameters was implemented. Correspondingly, the Analysis of Variance (ANOVA) has been employed to discuss the results taken from the developed model. Additionally, the essential parameters and interactions are extracted, and input parameters are sorted according to their importance. Moreover, the ANFIS using Takagi-Sugeno fuzzy system was employed. Finally, all methods were compared, and the effective parameters and associated relationships were extracted. In contrast to the other approaches, the ANFIS provided the best efficiency and high accuracy with the minimum desired errors. Comparatively, it was obtained that the ANN method is more effective than the RSM and has a higher regression coefficient and lower statistical errors.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.32 no.5
• /
• pp.817-826
• /
• 2022

AI technology is being successfully introduced in many fields, and models deployed as a service are deployed with black box environment that does not expose the model's information to protect intellectual property rights and data. In a black box environment, attackers try to steal data or parameters used during training by using model output. This paper proposes a method of inferring the type of model to directly find out the composition of layer of the target model, based on the fact that there is no attack to infer the information about the type of model from the deep learning model. With ResNet, VGGNet, AlexNet, and simple convolutional neural network models trained with MNIST datasets, we show that the types of models can be inferred using the output values in the gray box and black box environments of the each model. In addition, we inferred the type of model with approximately 83% accuracy in the black box environment if we train the big and small relationship feature that proposed in this paper together, the results show that the model type can be infrerred even in situations where only partial information is given to attackers, not raw probability vectors.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.365-381
• /
• 2023

The existing vision-based techniques for inspection and condition assessment of civil infrastructure are mostly manual and consequently time-consuming, expensive, subjective, and risky. As a viable alternative, researchers in the past resorted to deep learning-based autonomous damage detection algorithms for expedited post-disaster reconnaissance of structures. Although a number of automatic damage detection algorithms have been proposed, the scarcity of labeled training data remains a major concern. To address this issue, this study proposed a semi-supervised learning (SSL) framework based on consistency regularization and cross-supervision. Image data from post-earthquake reconnaissance, that contains cracks, spalling, and exposed rebars are used to evaluate the proposed solution. Experiments are carried out under different data partition protocols, and it is shown that the proposed SSL method can make use of unlabeled images to enhance the segmentation performance when limited amount of ground truth labels are provided. This study also proposes DeepLab-AASPP and modified versions of U-Net++ based on channel-wise attention mechanism to better segment the components and damage areas from images of reinforced concrete buildings. The channel-wise attention mechanism can effectively improve the performance of the network by dynamically scaling the feature maps so that the networks can focus on more informative feature maps in the concatenation layer. The proposed DeepLab-AASPP achieves the best performance on component segmentation and damage state segmentation tasks with mIoU scores of 0.9850 and 0.7032, respectively. For crack, spalling, and rebar segmentation tasks, modified U-Net++ obtains the best performance with Igou scores (excluding the background pixels) of 0.5449, 0.9375, and 0.5018, respectively. The proposed architectures win the second place in IC-SHM2021 competition in all five tasks of Project 2.

• Composites Research
• /
• v.37 no.1
• /
• pp.40-45
• /
• 2024

In this research, we introduce a novel approach that employs a 3D convolutional neural network (CNN) model to predict the permeability of Gas Diffusion Layers (GDLs). For training the model, we create an artificial dataset of GDL representative volume elements (RVEs) by extracting morphological characteristics from actual GDL images obtained through X-ray tomography. These morphological attributes involve statistical distributions of porosity, fiber orientation, and diameter. Subsequently, a permeability analysis using the Lattice Boltzmann Method (LBM) is conducted on a collection of 10,800 RVEs. The 3D CNN model, trained on this artificial dataset, well predicts the permeability of actual GDLs.

• Journal of Radiation Protection and Research
• /
• v.40 no.1
• /
• pp.55-64
• /
• 2015

Dose calculations which are a crucial requirement for radiotherapy treatment planning systems require accuracy and rapid calculations. The conventional radiotherapy treatment planning dose algorithms are rapid but lack precision. Monte Carlo methods are time consuming but the most accurate. The new combined system that Monte Carlo methods calculate part of interesting domain and the rest is calculated by neural can calculate the dose distribution rapidly and accurately. The preliminary study showed that neural networks can map functions which contain discontinuous points and inflection points which the dose distributions in inhomogeneous media also have. Performance results between scaled conjugated gradient algorithm and Levenberg-Marquardt algorithm which are used for training the neural network with a different number of neurons were compared. Finally, the dose distributions of homogeneous phantom calculated by a commercialized treatment planning system were used as training data of the neural network. In the case of homogeneous phantom;the mean squared error of percent depth dose was 0.00214. Further works are programmed to develop the neural network model for 3-dimensinal dose calculations in homogeneous phantoms and inhomogeneous phantoms.