• Journal of Korea Multimedia Society
• /
• v.23 no.2
• /
• pp.186-203
• /
• 2020

Image-based Virtual Try-On Systems are among the most potential solution for virtual fitting which tries on a target clothes into a model person image and thus have attracted considerable research efforts. In many cases, current solutions for those fails in achieving naturally looking virtual fitted image where a target clothes is transferred into the body area of a model person of any shape and pose while keeping clothes context like texture, text, logo without distortion and artifacts. In this paper, we propose a new improved image-based virtual try-on network system based on keypoints, which we name as KP-VTON. The proposed KP-VTON first detects keypoints in the target clothes and reliably predicts keypoints in the clothes of a model person image by utilizing a dense human pose estimation. Then, through TPS transformation calculated by utilizing the keypoints as control points, the warped target clothes image, which is matched into the body area for wearing the target clothes, is obtained. Finally, a new try-on module adopting Attention U-Net is applied to handle more detailed synthesis of virtual fitted image. Extensive experiments on a well-known dataset show that the proposed KP-VTON performs better the state-of-the-art virtual try-on systems.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.680-682
• /
• 2018

기후변화에 따라 매개 질병의 발병 빈도가 증가하고 있으며 모기와 같은 매개체에 의해 전염되는 매개 질병은 인구집단에 대한 중요한 위협 요인이다. 이러한 질병 관리를 위해 지역별 모기 서식 현황을 모니터링 하는 시스템의 필요성이 강조되고 있다. 하지만 현재의 모기 모니터링은 개체 파악을 위한 분류와 동정을 사람이 직접 수행하기에 오랜 시간이 소요된다. 이 연구는 그러한 문제점을 해결하고 미래 매개곤충 서식 현황 파악 시스템의 기반을 마련하기 위해 심층 신경망(Deep Neural Networks)을 활용하여 한국 주요 매개모기 종 분류를 수행하고 결과를 분석하였다. 종 분류를 위한 모델은 잘 알려진 신경망 모델인 DenseNet(Densely Connected Networks)을 사용하였고 이를 직접 촬영한 모기 데이터와 약간의 변형을 가한 모기 데이터를 사용하여 학습시켰다. 학습 데이터를 각각 5배, 20배, 100배로 증강하여 실제 데이터의 부족을 보완하였으며, 이를 통해 최대 99.48%의 정확도를 달성하였다.

• Nuclear Engineering and Technology
• /
• v.50 no.5
• /
• pp.724-730
• /
• 2018

Chromium and chromium nitride were selected as potential barriers to prevent fuel-clad chemical interaction (FCCI) between the cladding and the fuel material. In this study, ferritic/martensitic HT-9 steel and misch metal were used to simulate the reaction between the cladding and fuel fission product, respectively. Radio frequency magnetron sputtering was used to deposit Cr and CrN films onto the cladding, and the gas flow rates of argon and nitrogen were fixed at certain values for each sample to control the deposition rate and the crystal structure of the films. The samples were heated for 24 h at 933 K through the diffusion couple test, and considerable amount of interdiffusion (max. thickness: $550{\mu}m$) occurred at the interface between HT-9 and misch metal when the argon and nitrogen were used individually. The elemental contents of misch metal were detected at the HT-9 through energy dispersive X-ray spectroscopy due to the interdiffusion. However, the specimens that were sputtered by mixed gases (Ar and $N_2$) exhibited excellent resistance to FCCI. The thickness of these CrN films were only $4{\mu}m$, but these films effectively prevented the FCCI due to their high adhesion strength (frictional force ${\geq}1,200{\mu}m$) and dense columnar microstructures.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.44 no.3
• /
• pp.33-38
• /
• 2021

Recently, transfer learning techniques with a base convolutional neural network (CNN) model have widely gained acceptance in early detection and classification of crop diseases to increase agricultural productivity with reducing disease spread. The transfer learning techniques based classifiers generally achieve over 90% of classification accuracy for crop diseases using dataset of crop leaf images (e.g., PlantVillage dataset), but they have ability to classify only the pre-trained diseases. This paper provides with an evaluation scheme on selecting an effective base CNN model for crop disease transfer learning with regard to the accuracy of trained target crops as well as of untrained target crops. First, we present transfer learning models called CDC (crop disease classification) architecture including widely used base (pre-trained) CNN models. We evaluate each performance of seven base CNN models for four untrained crops. The results of performance evaluation show that the DenseNet201 is one of the best base CNN models.

• Nuclear Engineering and Technology
• /
• v.51 no.4
• /
• pp.1069-1074
• /
• 2019

A novel fabrication method of oxide dispersion strengthened (ODS) steel cladding tubes for advanced fast reactors has been investigated using the cold spray powder-based materials deposition process. Cold spraying has the potential advantage for rapidly fabricating ODS cladding tubes in comparison with the conventional multi-step extrusion process. A gas atomized spherical 14YWT (Fe-14%Cr, 3%W, 0.4%Ti, 0.2% Y, 0.01%O) powder was sprayed on a rotating cylindrical 6061-T6 aluminum mandrel using nitrogen as the propellant gas. The powder lacked the oxygen content needed to precipitate the nanoclusters in ODS steel, therefore this work was intended to serve as a proof-of-concept study to demonstrate that free-standing steel cladding tubes with prototypical ODS composition could be manufactured using the cold spray process. The spray process produced an approximately 1-mm thick, dense 14YWT deposit on the aluminum-alloy tube. After surface polishing of the 14YWT deposit to obtain desired cladding thickness and surface roughness, the aluminum-alloy mandrel was dissolved in an alkaline medium to leave behind a free-standing ODS tube. The as-fabricated cladding tube was annealed at $1000^{\circ}C$ for 1 h in an argon atmosphere to improve the overall mechanical properties of the cladding.

• Korean Journal of Remote Sensing
• /
• v.37 no.4
• /
• pp.747-761
• /
• 2021

With the increasing application of UAVs in intelligent transportation systems, vehicle detection for aerial images has become an essential engineering technology and has academic research significance. In this paper, a vehicle detection method for aerial images based on the YOLOv4 deep learning algorithm is presented. At present, the most known datasets are VOC (The PASCAL Visual Object Classes Challenge), ImageNet, and COCO (Microsoft Common Objects in Context), which comply with the vehicle detection from UAV. An integrated dataset not only reflects its quantity and photo quality but also its diversity which affects the detection accuracy. The method integrates three public aerial image datasets VAID, UAVD, DOTA suitable for YOLOv4. The training model presents good test results especially for small objects, rotating objects, as well as compact and dense objects, and meets the real-time detection requirements. For future work, we will integrate one more aerial image dataset acquired by our lab to increase the number and diversity of training samples, at the same time, while meeting the real-time requirements.

• Nuclear Engineering and Technology
• /
• v.54 no.7
• /
• pp.2650-2659
• /
• 2022

The dense granular flow spallation target is a new target concept proposed for an Accelerator-Driven Sub-critical (ADS) system. In this paper, the beam-target configurations of a tungsten granular flow target for the ADS with a thermal power of 1 GW is explored. The beam profile options using different scanning methods are discussed. The critical geometry parameters are adjusted to investigate the performance of the granular target from the aspects of neutron efficiency, stability and temperature distribution in target medium. To figure out how the target under accident conditions would behave, different clogging conditions are induced in the simulation. The dynamic processes are analyzed and some important parameters such as abnormal temperature rise and beam cutoff time window are obtained. The response of the sub-critical reactor to a clogging accident is also investigated. It is indicated that the monitoring of the granular flow by the neutron detectors in the sub-critical core will be effective.

• Nuclear Engineering and Technology
• /
• v.54 no.3
• /
• pp.991-996
• /
• 2022

Microstructures of monolithic high purity SiC and SiC with sintering additives after neutron irradiation to a fluence of 2.0-2.5 × 10²⁴ n/m² (E > 0.1 MeV) at 333-363 K and after post-irradiation annealing up to 1673 K were observed using a transmission electron microscopy. Results showed that no black spot defects or dislocation loops in SiC grains were found after the neutron irradiation for all of the specimens owing to the moderate fluence at low irradiation temperature. Thus, it is confirmed that these specimens were swelled mostly by the formation of point defects. Black spots and small dislocation loops were discovered only after the annealing process in PureBeta-SiC and CVD-SiC, where the swelling almost diminished. Anomalous-shaped YAG grains were found in SiC ceramics containing sintering additives. These grains contained dense black spots defects and might lose crystallinity after the neutron irradiation, while these defects may annihilate by recrystallization during annealing up to 1673 K. Amorphous grain boundary phase was also presented in this ceramic, and a large part of it was crystallized through post-irradiation annealing and could affect their recovery behavior.

• International Journal of Computer Science & Network Security
• /
• v.22 no.9
• /
• pp.23-30
• /
• 2022

Glaucoma is a chronic neuropathy that affects the optic nerve which can lead to blindness. The detection and prediction of glaucoma become possible using deep neural networks. However, the detection performance relies on the availability of a large number of data. Therefore, we propose different frameworks, including a hybrid of a generative adversarial network and a convolutional neural network to automate and increase the performance of glaucoma detection. The proposed frameworks are evaluated using five public glaucoma datasets. The framework which uses a Deconvolutional Generative Adversarial Network (DCGAN) and a DenseNet pre-trained model achieves 99.6%, 99.08%, 99.4%, 98.69%, and 92.95% of classification accuracy on RIMONE, Drishti-GS, ACRIMA, ORIGA-light, and HRF datasets respectively. Based on the experimental results and evaluation, the proposed framework closely competes with the state-of-the-art methods using the five public glaucoma datasets without requiring any manually preprocessing step.

• Journal of the Korean Geographical Society
• /
• v.41 no.2 s.113
• /
• pp.150-167
• /
• 2006

This study was conducted to determine the pedogenesis of dense subsurface horizons (denoted either Bx or Bd) observed within closed depressions and in toeslope positions at loess-covered glacial tillplains in southern Wisconsin. Some of these dense subsurface horizons, especially those occurring within depressions, show a close morphological resemblance to fragipans elsewhere, even though the existence of fragipans has not been previously reported in southern Wisconsin. The spatial occurrence of fragipans was first examined over the landscape to characterize general soil-landscape relationships. Detailed physico-chemical and micromorphological analyses were followed to investigate the development of fragipans within a closed depression along a catenary sequence. The formation of fragipans at the study site is a result of sequential processes of physical ripening and accumulation of colloidal materials. A very coarse prismatic structure with a closely packed soil matrix was formed via physical ripening processes of loess deposited in small glacial lakes and floodplains that existed soon after the retreat of the last glacier. The physically formed dense horizons became hardened by the accumulation of colloidal materials, notably amorphous Si. The accumulation intensity of amorphous Si varies with mass balance relationships, which are governed by topography and local drainage conditions. Well-developed Bx horizons evolve at closed depressions where net accumulation of amorphous Si occurs, but the collapsed layers remain as Bd horizons at other locations where soluble Si has continuously been removed downslope or downvalley. Hydromorphic processes caused by the presence of fragipans are degrading upper parts of the prisms, resulting in the formation of an eluvial fragic horizon (Ex).