• 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.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_2
• /
• pp.793-810
• /
• 2022

Synthetic Aperture Radar (SAR) is considered to be suitable for near real-time inundation monitoring. The distinctly different intensity between water and land makes it adequate for waterbody detection, but the intrinsic speckle noise and variable intensity of SAR images decrease the accuracy of waterbody detection. In this study, we suggest two modules, named 'morphology module' and 'edge-enhanced module', which are the combinations of pooling layers and convolutional layers, improving the accuracy of waterbody detection. The morphology module is composed of min-pooling layers and max-pooling layers, which shows the effect of morphological transformation. The edge-enhanced module is composed of convolution layers, which has the fixed weights of the traditional edge detection algorithm. After comparing the accuracy of various versions of each module for U-Net, we found that the optimal combination is the case that the morphology module of min-pooling and successive layers of min-pooling and max-pooling, and the edge-enhanced module of Scharr filter were the inputs of conv9. This morphologic and edge-enhanced U-Net improved the F1-score by 9.81% than the original U-Net. Qualitative inspection showed that our model has capability of detecting small-sized waterbody and detailed edge of water, which are the distinct advancement of the model presented in this research, compared to the original U-Net.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_1
• /
• pp.991-1005
• /
• 2022

Satellite imaging is an effective supplementary data source for detecting and verifying nuclear activity. It is also highly beneficial in regions with limited access and information, such as nuclear installations. Time series analysis, in particular, can identify the process of preparing for the conduction of a nuclear experiment, such as relocating equipment or changing facilities. Differences in the semantic segmentation findings of time series photos were employed in this work to detect changes in meaningful items connected to nuclear activity. Building, road, and small object datasets made of KOMPSAT 3/3A photos given by AIHub were used to train deep learning models such as U-Net, PSPNet, and Attention U-Net. To pick relevant models for targets, many model parameters were adjusted. The final change detection was carried out by including object information into the first change detection, which was obtained as the difference in semantic segmentation findings. The experiment findings demonstrated that the suggested approach could effectively identify altered pixels. Although the suggested approach is dependent on the accuracy of semantic segmentation findings, it is envisaged that as the dataset for the region of interest grows in the future, so will the relevant scope of the proposed method.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.40 no.5
• /
• pp.535-546
• /
• 2020

Land cover maps are derived from satellite and aerial images by the Ministry of Environment for the entire Korea since 1998. Even with their wide application in many sectors, their usage in research community is limited. The main reason for this is the map compilation cycle varies too much over the different regions. The situation requires us a new and quicker methodology for generating land cover maps. This study was conducted to automatically generate land cover map using aerial ortho-images and Landsat 8 satellite images. The input aerial and Landsat 8 image data were trained by Residual U-Net, one of the deep learning-based segmentation techniques. Study was carried out by dividing three groups. First and second group include part of level-II (medium) categories and third uses group level-III (large) classification category defined in land cover map. In the first group, the results using all 7 classes showed 86.6 % of classification accuracy The other two groups, which include level-II class, showed 71 % of classification accuracy. Based on the results of the study, the deep learning-based research for generating automatic level-III classification was presented.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.5
• /
• pp.323-330
• /
• 2023

The development of an analysis model that reflects the microstructure characteristics of polyvinyl alcohol (PVA) fiber-reinforced cementitious composites, which have a highly complex microstructure, enables synergy between efficient material design and real experiments. PVA fiber orientations are an important factor that influences the mechanical behavior of PVA fiber-reinforced cementitious composites. Owing to the difficulty in distinguishing the gray level value obtained from micro-CT images of PVA fibers from adjacent phases, fiber segmentation is time-consuming work. In this study, a micro-CT test with a voxel size of 0.65 ㎛³ was performed to investigate the three-dimensional distribution of fibers. To segment the fibers and generate training data, histogram, morphology, and gradient-based phase-segmentation methods were used. A U-net model was proposed to segment fibers from micro-CT images of PVA fiber-reinforced cementitious composites. Data augmentation was applied to increase the accuracy of the training, using a total of 1024 images as training data. The performance of the model was evaluated using accuracy, precision, recall, and F1 score. The trained model achieved a high fiber segmentation performance and efficiency, and the approach can be applied to other specimens as well.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.27 no.2
• /
• pp.49-70
• /
• 2022

Sea Surface Temperature (SST), one of the ocean features, has a significant impact on climate, marine ecosystem and human activities. Therefore, SST prediction has been always an important issue. Recently, deep learning has drawn much attentions, since it can predict SST by training past SST patterns. Compared to the numerical simulations, deep learning model is highly efficient, since it can estimate nonlinear relationships between input data. With the recent development of Graphics Processing Unit (GPU) in computer, large amounts of data can be calculated repeatedly and rapidly. In this study, Short-term SST will be predicted through Convolutional Neural Network (CNN)-based U-Net that can handle spatiotemporal data concurrently and overcome the drawbacks of previously existing deep learning-based models. The SST prediction performance depends on the seasonal and interannual SST variabilities around the southern coast of Korea. The predicted SST has a wide range of variance during spring and summer, while it has small range of variance during fall and winter. A wide range of variance also has a significant correlation with the change of the Pacific Decadal Oscillation (PDO) index. These results are found to be affected by the intensity of the seasonal and PDO-related interannual SST fronts and their intensity variations along the southern Korean seas. This study implies that the SST prediction performance using the developed deep learning model can be significantly varied by seasonal and interannual variabilities in SST.

• KIPS Transactions on Software and Data Engineering
• /
• v.13 no.1
• /
• pp.17-27
• /
• 2024

Video frame interpolation is an important technique used in the field of video and media, as it increases the continuity of motion and enables smooth playback of videos. In the study of video frame interpolation using deep learning, Kernel Based Method captures local changes well, but has limitations in handling global changes. In this paper, we propose a new U-Net structure that applies feature map differentiation and two directions to focus on capturing major changes to generate intermediate frames more accurately while reducing the number of parameters. Experimental results show that the proposed structure outperforms the existing model by up to 0.3 in PSNR with about 61% fewer parameters on common datasets such as Vimeo, Middle-burry, and a new YouTube dataset. Code is available at https://github.com/Go-MinSeong/SF-AdaCoF.

• Journal of the Korean Society of Visualization
• /
• v.20 no.3
• /
• pp.109-116
• /
• 2022

The optimized design of airfoils is essential to increase the performance and efficiency of wind turbines. The aerodynamic characteristics of airfoils near the stall show large deviation from experiments and numerical simulations. Hence, it is needed to perform repetitive analysis of various shapes near the stall. To overcome this, the artificial intelligence is used and combined with numerical simulations. In this study, three types of airfoils are chosen, which are S809, S822 and SD7062 used in wind turbines. A convolutional neural network model is proposed in the combination of VGG16 and U-Net. Learning data are constructed by extracting pressure fields and aerodynamic characteristics through numerical analysis of 2D shape. Based on these data, the pressure field and lift coefficient of untrained airfoils are predicted. As a result, even in untrained airfoils, the pressure field is accurately predicted with an error of within 0.04%.

• Smart Media Journal
• /
• v.13 no.4
• /
• pp.86-93
• /
• 2024

The development of functional magnetic resonance imaging (fMRI) has significantly contributed to mapping brain functions and understanding brain networks during rest. This paper proposes a CNN-LSTM-based classification model to classify the progression stages of Alzheimer's disease. Firstly, four preprocessing steps are performed to remove noise from the fMRI data before feature extraction. Secondly, the U-Net architecture is utilized to extract spatial features once preprocessing is completed. Thirdly, the extracted spatial features undergo LSTM processing to extract temporal features, ultimately leading to classification. Experiments were conducted by adjusting the temporal dimension of the data. Using 5-fold cross-validation, an average accuracy of 96.4% was achieved, indicating that the proposed method has high potential for identifying the progression of Alzheimer's disease by analyzing fMRI data.