• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.68 no.2
• /
• pp.359-363
• /
• 2019

As commercial drones have been widely used, concerns for collision accidents with people and invading secured properties are emerging. The detection of drone is a challenging problem. The deep learning based object detection techniques for detecting drones have been applied, but limited to the specific cases such as detection of drones from bird and/or background. We have tried not only detection of drones, but classification of different drones with an end-to-end model. YOLOv2 is used as an object detection model. In order to supplement insufficient data by shooting drones, data augmentation from collected images is executed. Also transfer learning from ImageNet for YOLOv2 darknet framework is performed. The experimental results for drone detection with average IoU and recall are compared and analysed.

• Nuclear Engineering and Technology
• /
• v.53 no.3
• /
• pp.866-878
• /
• 2021

Procedures play a key role in ensuring safe operation at nuclear power plants (NPPs). Development and maintenance of a large number of procedures reflecting the best knowledge available in all relevant areas is a complex job. This paper introduces a newly developed methodology and the implemented software, called iExtractor, for the extraction of syntactic and semantic information from NPP procedures utilizing natural language processing (NLP)-based technologies. The steps of the iExtractor integrated with sets of rules and an ontology for NPPs are described in detail with examples. Case study results of the iExtractor applied to selected procedures of a U.S. commercial NPP are also introduced. It is shown that the iExtractor can provide overall comprehension of the analyzed procedures and indicate parts of procedures that need improvement. The rich information extracted from procedures could be further utilized as a basis for their enhanced management.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2191-2208
• /
• 2022

Motion blur in PET (Positron emission tomography) images induced by respiratory motion will reduce the quality of imaging. Although exiting methods have positive performance for respiratory motion correction in medical practice, there are still many aspects that can be improved. In this paper, an improved 3D unsupervised framework, Res-Voxel based on U-Net network was proposed for the motion correction. The Res-Voxel with multiple residual structure may improve the ability of predicting deformation field, and use a smaller convolution kernel to reduce the parameters of the model and decrease the amount of computation required. The proposed is tested on the simulated PET imaging data and the clinical data. Experimental results demonstrate that the proposed achieved Dice indices 93.81%, 81.75% and 75.10% on the simulated geometric phantom data, voxel phantom data and the clinical data respectively. It is demonstrated that the proposed method can improve the registration and correction performance of PET image.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.60-79
• /
• 2022

The U-Net architecture-based segmentation models attained remarkable performance in numerous medical image segmentation missions like skin lesion segmentation. Nevertheless, the resolution gradually decreases and the loss of spatial information increases with deeper network. The fusion of adjacent layers is not enough to make up for the lost spatial information, thus resulting in errors of segmentation boundary so as to decline the accuracy of segmentation. To tackle the issue, we propose a new deep learning-based segmentation model. In the decoding stage, the feature channels of each decoding unit are concatenated with all the feature channels of the upper coding unit. Which is done in order to ensure the segmentation effect by integrating spatial and semantic information, and promotes the robustness and generalization of our model by combining the atrous spatial pyramid pooling (ASPP) module and channel attention module (CAM). Extensive experiments on ISIC2016 and ISIC2017 common datasets proved that our model implements well and outperforms compared segmentation models for skin lesion segmentation.

• Nuclear Engineering and Technology
• /
• v.54 no.9
• /
• pp.3260-3271
• /
• 2022

Experience shows that the solution to waste management in any national programme is lengthy and burdened with uncertainties. There are several uncertainties that contribute to the costs associated with spent fuel management. In this work, we have analysed the impact of the current nuclear data on the isotopic composition of the spent fuel and consequently their influence on the main spent fuel observables such as decay heat, activity, neutron multiplication factor, and neutron and photon source terms. Nuclear libraries based on the most general nuclear data ENDF/B-VII.0, ENDF/B-VII.1, ENDF/B-VIII.0 and JEFF-3.3 are considered. A typical NPP Krško fuel assembly is analysed using the Monte Carlo code Serpent 2. The analysis considers burnup of up to 60 GWd/tU and cooling times of up to 100 years. The comparison of results showed significant differences, which should be taken into account when selecting the library and evaluating the uncertainty in determining the characteristics of the spent fuel.

• Animal Systematics, Evolution and Diversity
• /
• v.22 no.2
• /
• pp.169-177
• /
• 2006

Some hyromedusae were collected from Korean waters with Issacc-Kidd Midwater trawl net or by SCUBA diving during the period from May 2001 to Nov. 2005. They were identified into 14 species of 11 families in six orders. Of these, the following six species are new to Korean fauna: Olindias formosa (Goto, 1903) and Proboscidactyla stellata (Forbes, 1846) of the order Limnomedusae; Aglantha digitale ($M\ddot{u}ller$, 1776) and Rhopalonema velatum Gegenbaur, 1854 of the order Trachymedusae; Solmundella bitentaculata (Quoy and Gaimaud, 1833) and Aegina citrea Eschscholtz, 1829 of the order Narcomedusae. The order Narcomedusae and the families Rhopalonematidae and Aeginidae are first recorded in Korean waters. As a result of the present study, the Korean hydromedusae consist of 26 species of 19 families in six orders so far.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.257-260
• /
• 2020

본 논문은 단일의 옷감 이미지로 가상의 그래픽 렌더링을 위해 Pix2Pix 방법을 이용하여 Normal map 을 생성하는 방법을 제시한다. 구체적으로 단일의 이미지를 이용해서 Normal map 를 생성하기 위해, Color image 와 Normal map 쌍의 training dataset 을 Pix2Pix 방법을 이용해서 학습시킨다 또한, test dataset 의 Color image 를 입력으로 넣어 생성된 Normal map 결과를 확인한다. 그리고 선행연구에서 사용되어오던 U-Net 방식의 방법과 본 논문에서 사용한 Pix2Pix 를 이용한 Normal map 생성 결과를 SSIM(Structural Similarity Index)으로 비교 평가한다. 또한, 생성된 Normal map 을 렌더링하고자 하는 가상 객체의 사이즈에 맞게 사이즈를 조정하여 OpenGL 로 렌더링한 결과를 확인한다. 본 논문을 통해서 단일의 패턴 이미지를 Pix2Pix 로 생성한 Normal map 으로 옷감의 디테일을 사실감 있게 표현할 수 있음을 확인할 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.13-14
• /
• 2023

뇌의 전기적 신경활동을 측정하는 뇌전도(EEG)는 저렴하게 취득할 수 있고 높은 시간 해상도를 갖는 반면 공간적 정보를 제공하지는 않는다. 기능적 자기공명영상(fMRI)은 혈류변화를 감지하여 뇌활동을 측정하는 방식으로서 높은 공간 분해능을 갖지만 고가의 비용과 설비를 요구한다. 최근 저렴하게 취득할 수 있는 EEG 데이터로부터 딥러닝을 사용하여 fMRI 합성영상을 생성하는 기술이 제안되었지만, 저주파수 대역에서 EEG와 fMRI 간의 뇌과학적 상관관계를 반영하지는 않는다. 본 연구에서는 휴식상태에서 취득된 EEG 데이터를 스펙트로그램으로 변환한 후 저주파수 특성을 사용하여 fMRI 합성영상을 생성하는 U-net 기반의 크로스 모달리티 변환 모델의 실현가능성을 평가하였다.

• Smart Structures and Systems
• /
• v.32 no.4
• /
• pp.219-233
• /
• 2023

Crack width is an important indicator to evaluate the health condition of the concrete structure. The crack width is measured by manual using crack width gauge commonly, which is time-consuming and laborious. In this paper, we have proposed a fast and simplified crack width quantification method via deep Q learning and geometric calculation. Firstly, the crack edge is extracted by using U-Net network and edge detection operator. Then, the intelligent decision of is made by the deep Q learning model. Further, the geometric calculation method based on endpoint and curvature extreme point detection is proposed. Finally, a case study is carried out to demonstrate the effectiveness of the proposed method, achieving high precision in the real crack width quantification.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.134-141
• /
• 2023

We present a novel method aimed at refining ground truth data through regularization and modification, particularly applicable when working with the original ground truth set. Enhancing the performance of deep neural networks is achieved by applying regularization techniques to the existing ground truth data. In many machine learning tasks requiring pixel-level segmentation sets, accurately delineating objects is vital. However, it proves challenging for thin and elongated objects such as blood vessels in X-ray coronary angiography, often resulting in inconsistent generation of ground truth data. This method involves an analysis of the quality of training set pairs - comprising images and ground truth data - to automatically regulate and modify the boundaries of ground truth segmentation. Employing the active contour model and a recursive ground truth generation approach results in stable and precisely defined boundary contours. Following the regularization and adjustment of the ground truth set, there is a substantial improvement in the performance of deep neural networks.