• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.103-110
• /
• 2022

Since satellite images generally include clouds in the atmosphere, it is essential to detect or mask clouds before satellite image processing. Clouds were detected using physical characteristics of clouds in previous research. Cloud detection methods using deep learning techniques such as CNN or the modified U-Net in image segmentation field have been studied recently. Since image segmentation is the process of assigning a label to every pixel in an image, precise pixel-based dataset is required for cloud detection. Obtaining accurate training datasets is more important than a network configuration in image segmentation for cloud detection. Existing deep learning techniques used different training datasets. And test datasets were extracted from intra-dataset which were acquired by same sensor and procedure as training dataset. Different datasets make it difficult to determine which network shows a better overall performance. To verify the effectiveness of the cloud detection network such as Cloud-Net, two types of networks were trained using the cloud dataset from KOMPSAT-3 images provided by the AIHUB site and the L8-Cloud dataset from Landsat8 images which was publicly opened by a Cloud-Net author. Test data from intra-dataset of KOMPSAT-3 cloud dataset were used for validating the network. The simulation results show that the network trained with KOMPSAT-3 cloud dataset shows good performance on the network trained with L8-Cloud dataset. Because Landsat8 and KOMPSAT-3 satellite images have different GSDs, making it difficult to achieve good results from cross-sensor validation. The network could be superior for intra-dataset, but it could be inferior for cross-sensor data. It is necessary to study techniques that show good results in cross-senor validation dataset in the future.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.708-710
• /
• 2021

본 논문에서는 X-ray 진단에서 산란선으로 인한 영상의 왜곡을 보정하는 방법으로서 수정된 구조의 AE(Auto-Encoder) 모델에 기반한 방법론을 제안한다. 기존 AE 모델의 계층에 따라 특징지도의 크기가 축소되고 팽창되는 과정에서 영상 복원에 필요한 정보가 소실될 가능성을 보완하기 위하여 동일 레벨 계층 간에 스킵 연결을 추가하였다. 또한 X-ray 영상에서 피사체 세부 부위의 두께와 밀도에 따라 산란선의 영향이 서로 다른 형태로 나타난다는 특성을 학습 과정에 효과적으로 반영하기 위하여 어텐션 모듈을 추가한 네트워크 구조를 도입하였다. 총 80 쌍의 흉부 X-ray 영상 데이터에 대하여 기존의 AE 모델을 사용한 방법 및 U-Net 과 FFA-Net 모델을 사용한 영상 복원 기법의 실험 결과를 상호 비교함으로써 제안된 방법의 타당성을 평가하였다.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.9
• /
• pp.99-112
• /
• 2019

Cracks on pavement surfaces are critical signs and symptoms of the degradation of pavement structures. Image-based pavement crack detection is a challenging problem due to the intensity inhomogeneity, topology complexity, low contrast, and noisy texture background. In this paper, we address the problem of pavement crack detection and segmentation at pixel-level based on a Deep Neural Network (DNN) using gray-scale images. We propose a novel DNN architecture which contains a modified U-net network and a high-level features network. An important contribution of this work is the combination of these networks afforded through the fusion layer. To the best of our knowledge, this is the first paper introducing this combination for pavement crack segmentation and detection problem. The system performance of crack detection and segmentation is enhanced dramatically by using our novel architecture. We thoroughly implement and evaluate our proposed system on two open data sets: the Crack Forest Dataset (CFD) and the AigleRN dataset. Experimental results demonstrate that our system outperforms eight state-of-the-art methods on the same data sets.

• Nuclear Engineering and Technology
• /
• v.45 no.7
• /
• pp.827-838
• /
• 2013

Interaction layer growth between U-Mo alloy fuel particles and Al in a dispersion fuel is a concern due to the volume expansion and other unfavorable irradiation behavior of the interaction product. To reduce interaction layer (IL) growth, a small amount of Si is added to the Al. As a result, IL growth is affected by the Si content in the Al matrix. In order to predict IL growth during fabrication and irradiation, empirical models were developed. For IL growth prediction during fabrication and any follow-on heating process before irradiation, out-of-pile heating test data were used to develop kinetic correlations. Two out-of-pile correlations, one for the pure Al matrix and the other for the Al matrix with Si addition, respectively, were developed, which are Arrhenius equations that include temperature and time. For IL growth predictions during irradiation, the out-of-pile correlations were modified to include a fission-rate term to consider fission enhanced diffusion, and multiplication factors to incorporate the Si addition effect and the effect of the Mo content. The in-pile correlation is applicable for a pure Al matrix and an Al matrix with the Si content up to 8 wt%, for fuel temperatures up to $200^{\circ}C$, and for Mo content in the range of 6 - 10wt%. In order to cover these ranges, in-pile data were included in modeling from various tests, such as the US RERTR-4, -5, -6, -7 and -9 tests and Korea's KOMO-4 test, that were designed to systematically examine the effects of the fission rate, temperature, Si content in Al matrix, and Mo content in U-Mo particles. A model converting the IL thickness to the IL volume fraction in the meat was also developed.

• Journal of ILASS-Korea
• /
• v.9 no.2
• /
• pp.18-23
• /
• 2004

An experimental studies of conventional gasoline fuel pump were carried out to obtain fundamental data fur liquid phase LPG injection(LPLi) system. A regenerative type and a roller-vane type of pumps were investigated in various operational condition. The experiments were performed to obtain flow rate of LPG fuel as a function of pressure differences and temperatures. The regenerative pump had too low flow rate at some experimental conditions to use this pump system for LPLi fuel supply system. On the other hand, the roller-vane type pump can be applied to the system only if its check valve is modified. Cavitation might occur in this system which can result in system noise, flow rate variation, and pump durability problem. To solve these problems the system is needed to increase $NPSH_{re}$(required net positive suction head).

• Nuclear Engineering and Technology
• /
• v.55 no.10
• /
• pp.3913-3923
• /
• 2023

Todays, medium energy resolution detectors are preferably used in radioisotope identification devices(RID) in nuclear and radioactive material categorization. However, there is still a need to develop or enhance « automated identifiers » for the useful RID algorithms. To decide whether any material is SNM or NORM, a key parameter is the better energy resolution of the detector. Although masking, shielding and gain shift/stabilization and other affecting parameters on site are also important for successful operations, the suitability of the RID algorithm is also a critical point to enhance the identification reliability while extracting the features from the spectral analysis. In this study, a RID algorithm based on Bayesian statistical method has been modified for medium energy resolution detectors and applied to the uranium gamma-ray spectra taken by a LaBr3:Ce detector. The present Bayesian RID algorithm covers up to 2000 keV energy range. It uses the peak centroids, the peak areas from the measured gamma-ray spectra. The extraction features are derived from the peak-based Bayesian classifiers to estimate a posterior probability for each isotope in the ANSI library. The program operations were tested under a MATLAB platform. The present peak based Bayesian RID algorithm was validated by using single isotopes(²⁴¹Am, ⁵⁷Co, ¹³⁷Cs, ⁵⁴Mn, ⁶⁰Co), and then applied to five standard nuclear materials(0.32-4.51% at.²³⁵U), as well as natural U- and Th-ores. The ID performance of the RID algorithm was quantified in terms of F-score for each isotope. The posterior probability is calculated to be 54.5-74.4% for ²³⁸U and 4.7-10.5% for ²³⁵U in EC-NRM171 uranium materials. For the case of the more complex gamma-ray spectra from CRMs, the total scoring (ST) method was preferred for its ID performance evaluation. It was shown that the present peak based Bayesian RID algorithm can be applied to identify ²³⁵U and ²³⁸U isotopes in LEU or natural U-Th samples if a medium energy resolution detector is was in the measurements.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.1
• /
• pp.1-7
• /
• 2023

An artificial intelligence (AI) method based on image deep learning is proposed to predict the entire displacement shape of a structure using the feature of partial displacements. The performance of the method was investigated through a structural test of a steel frame. An image-to-image regression (I2IR) training method was developed based on the U-Net layer for image recognition. In the I2IR method, the U-Net is modified to generate images of entire displacement shapes when images of partial displacement shapes of structures are input to the AI network. Furthermore, the training of displacements combined with the location feature was developed so that nodal displacement values with corresponding nodal coordinates could be used in AI training. The proposed training methods can consider correlations between nodal displacements in 3D space, and the accuracy of displacement predictions is improved compared with artificial neural network training methods. Displacements of the steel frame were predicted during the structural tests using the proposed methods and compared with 3D scanning data of displacement shapes. The results show that the proposed AI prediction properly follows the measured displacements using 3D scanning.

• Journal of the East Asian Society of Dietary Life
• /
• v.9 no.3
• /
• pp.378-385
• /
• 1999

A reduced NaCl intake for the general population of the world has been recommended to reduce the overall blood pressure level and hence to reduce the overall incidence of cardiovascular disease. A high NaCl diet convincingly contributes to elevated arterial pressure in humans and animal models of hypertension. Among individuals there is considerable variability of blood pressure responsiveness to NaCl intake. In normotensive as well as hypertensive subjects, blood pressure can be judged to be salt sensitivity (SS) when observed to vary directly and substantially with the net intake of NaCl. The prevalence of SS in normotensive adults in the U.S. ranges from 15% to 42% and in hypertensive adults from 28% to 74%. SS is a risk factor for hypertension and may be an important marker in the identification of children for hypertension prevention programs. High NaCl intakes produce expansion of the extracellular fluid volume and thus increase blood pressure. Nonchloride salts of sodium does not expand the extracellular fluid volume and does not alter blood pressure. Blood pressure response to NaCl may be modified by other components of the diet. Low dietary intakes of potassium or calcium augment NaCl-induced increases of blood pressure. Conversely, high dietary intakes of potassium or calcium attenuate NaCl-induced hypertension. A greater intakes of potassium or calcium may prevent or delay the occurrence of hypertension. SS occurs when dietary potassium is even marginally deficient but is dose-dependently suppressed when dietary potassium is increased within its normal range. Orally administered KHCO$_3$, abundant in fruits and vegetates, but not KCl has a calcium-retaining effect which may contributed to its reversal of pressor effect of dietary NaCl. Since nutrients other than NaCl also affect blood pressure levels, a reduced NaCl intake should be only one component of a nutritional strategy to lower blood pressure.

• Nuclear Engineering and Technology
• /
• v.47 no.7
• /
• pp.791-803
• /
• 2015

The effective cross sections (XSs) in the direct whole core calculation code nTRACER are evaluated by the equivalence theory-based resonance-integral-table method using the WIMS-based library as an alternative to the subgroup method. The background XSs, as well as the Dancoff correction factors, were evaluated by the enhanced neutron-current method. A method, with pointwise microscopic XSs on a union-lethargy grid, was used for the generation of resonance-interference factors (RIFs) for mixed resonant absorbers. This method was modified by the intermediate-resonance approximation by replacing the potential XSs for the non-absorbing moderator nuclides with the background XSs and neglecting the resonance-elastic scattering. The resonance-escape probability was implemented to incorporate the energy self-shielding effect in the spectrum. The XSs were improved using the proposed method as compared to the narrow resonance infinite massbased method. The RIFs were improved by 1% in $^{235}U$, 7% in $^{239}Pu$, and >2% in $^{240}Pu$. To account for thermal feedback, a new feature was incorporated with the interpolation of pre-generated RIFs at the multigroup level and the results compared with the conventional resonance-interference model. This method provided adequate results in terms of XSs and k-eff. The results were verified first by the comparison of RIFs with the exact RIFs, and then comparing the XSs with the McCARD calculations for the homogeneous configurations, with burned fuel containing a mixture of resonant nuclides at different burnups and temperatures. The RIFs and XSs for the mixture showed good agreement, which verified the accuracy of the RIF evaluation using the proposed method. The method was then verified by comparing the XSs for the virtual environment for reactor applicationbenchmark pin-cell problem, as well as the heterogeneous pin cell containing burned fuel with McCARD. The method works well for homogeneous, as well as heterogeneous configurations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.881-895
• /
• 2023

Modern image inpainting techniques based on deep learning have achieved remarkable performance, and more and more people are working on repairing more complex and larger missing areas, although this is still challenging, especially for facial image inpainting. For a face image with a huge missing area, there are very few valid pixels available; however, people have an ability to imagine the complete picture in their mind according to their subjective will. It is important to simulate this capability while maintaining the identity features of the face as much as possible. To achieve this goal, we propose a three-stage network model, which we refer to as the identity and structure feature refinement network (ISFRNet). ISFRNet is based on 1) a pre-trained pSp-styleGAN model that generates an extremely realistic face image with rich structural features; 2) a shallow structured network with a small receptive field; and 3) a modified U-net with two encoders and a decoder, which has a large receptive field. We choose structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), L1 Loss and learned perceptual image patch similarity (LPIPS) to evaluate our model. When the missing region is 20%-40%, the above four metric scores of our model are 28.12, 0.942, 0.015 and 0.090, respectively. When the lost area is between 40% and 60%, the metric scores are 23.31, 0.840, 0.053 and 0.177, respectively. Our inpainting network not only guarantees excellent face identity feature recovery but also exhibits state-of-the-art performance compared to other multi-stage refinement models.