• Journal of architectural history
• /
• v.31 no.1
• /
• pp.41-50
• /
• 2022

Traditional wooden buildings deform over time and are vulnerable to fire or earthquakes. Therefore, traditional wooden buildings require continuous management and repair, and securing architectural drawings is essential for repair and restoration. Unlike modernized CAD drawings, traditional wooden building drawings scan and store hand-drawn drawings, and in this process, many noise is included due to damage to the drawing itself. These drawings are digitized, but their utilization is poor due to noise. Difficulties in systematic management of traditional wooden buildings are increasing. Noise removal by existing algorithms has limited drawings that can be applied according to noise characteristics and the performance is not uniform. This study presents deep artificial neural network based noised reduction for architectural drawings. Front/side elevation drawings, floor plans, detail drawings of Korean wooden treasure buildings were considered. First, the noise properties of the architectural drawings were learned with both a cycle generative model and heuristic image fusion methods. Consequently, a noise reduction network was trained through supervised learning using training sets prepared using the noise models. The proposed method provided effective removal of noise without deteriorating fine lines in the architectural drawings and it showed good performance for various noise types.

• The Journal of Bigdata
• /
• v.7 no.1
• /
• pp.173-185
• /
• 2022

Nowadays, Korean webtoons are leading the global digital comic market. Webtoons are being serviced in various languages around the world, and dramas or movies produced with Webtoons' IP (Intellectual Property Rights) have become a big hit, and more and more webtoons are being visualized. However, with the success of these webtoons, the working environment of webtoon creators is emerging as an important issue. According to the 2021 Cartoon User Survey, webtoon creators spend 10.5 hours a day on creative activities on average. Creators have to draw large amount of pictures every week, and competition among webtoons is getting fiercer, and the amount of paintings that creators have to draw per episode is increasing. Therefore, this study proposes to generate webtoon background images using deep learning algorithms and use them for webtoon production. The main character in webtoon is an area that needs much of the originality of the creator, but the background picture is relatively repetitive and does not require originality, so it can be useful for webtoon production if it can create a background picture similar to the creator's drawing style. Background generation uses CycleGAN, which shows good performance in image-to-image translation, and CartoonGAN, which is specialized in the Cartoon style image generation. This deep learning-based image generation is expected to shorten the working hours of creators in an excessive work environment and contribute to the convergence of webtoons and technologies.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.11 no.2
• /
• pp.115-131
• /
• 2013

The Mizunami Underground Laboratory (MIU) Project is a comprehensive research project investigating the deep underground environment within crystalline rock being conducted by Japan Atomic Energy Agency. The MIU Project has three overlapping phases: Surface-based Investigation phase (Phase I), Construction phase (Phase II), and Operation phase (Phase III), with a total duration of 20 years. The overall project goals of the MIU Project from Phase I through to Phase III are: 1) to establish techniques for investigation, analysis and assessment of the deep geological environment, and 2) to develop a range of engineering for deep underground application. For the overall project goals 1), the Phase I goals were set to construct models of the geological environment from all surface-based investigation results that describe the geological environment prior to excavation and predict excavation response. For the overall project goals 2), the Phase I goals were set to formulate detailed design concepts and a construction plan for the underground facilities. This paper introduces geosynthesis procedures for the investigation and assessment of the hydrochemistry of groundwater in crystalline rock.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2016.02a
• /
• pp.310-310
• /
• 2016

최근 NAND flash memory는 높은 집적성과 데이터의 비휘발성, 낮은 소비전력, 간단한 입, 출력 등의 장점들로 인해 핸드폰, MP3, USB 등의 휴대용 저장 장치 및 노트북 시장에서 많이 이용되어 왔다. 특히, 최근에는 smart watch, wearable device등과 같은 차세대 디스플레이 소자에 대한 관심이 증가함에 따라 유연하고 투명한 메모리 소자에 대한 연구가 다양하게 진행되고 있다. 대표적인 플래시 메모리 소자의 구조로 charge trapping type flash memory (CTF)가 있다. CTF 메모리 소자는 trap layer의 trap site를 이용하여 메모리 동작을 하는 소자이다. 하지만 작은 window의 크기, trap site의 열화로 인해 메모리 특성이 나빠지는 문제점 등이 있다. 따라서 최근, trap layer에 다양한 물질을 적용하여 CTF 소자의 문제점을 해결하고자 하는 연구들이 진행되고 있다. 특히, 산화물 반도체인 zinc oxide (ZnO)를 trap layer로 하는 CTF 메모리 소자가 최근 몇몇 보고 되었다. 산화물 반도체인 ZnO는 n-type 반도체이며, shallow와 deep trap site를 동시에 가지고 있는 독특한 물질이다. 이 특성으로 인해 메모리 소자의 programming 시에는 deep trap site에 charging이 일어나고, erasing 시에는 shallow trap site에 캐리어들이 쉽게 공급되면서 deep trap site에 갇혀있던 charge가 쉽게 de-trapped 된다는 장점을 가지고 있다. 따라서, 본 실험에서는 산화물 반도체인 ZnO를 trap layer로 하는 CTF 소자의 메모리 특성을 확인하기 위해 간단한 구조인 metal-oxide capacitor (MOSCAP)구조로 제작하여 메모리 특성을 평가하였다. 먼저, RCA cleaning 처리된 n-Si bulk 기판 위에 tunnel layer인 SiO2 5 nm를 rf sputter로 증착한 후 furnace 장비를 이용하여 forming gas annealing을 $450^{\circ}C$에서 실시하였다. 그 후 ZnO를 20 nm, SiO2를 30 nm rf sputter로 증착한 후, 상부전극을 E-beam evaporator 장비를 사용하여 Al 150 nm를 증착하였다. 제작된 소자의 신뢰성 및 내구성 평가를 위해 상온에서 retention과 endurance 측정을 진행하였다. 상온에서의 endurance 측정결과 1000 cycles에서 약 19.08%의 charge loss를 보였으며, Retention 측정결과, 10년 후 약 33.57%의 charge loss를 보여 좋은 메모리 특성을 가지는 것을 확인하였다. 본 실험 결과를 바탕으로, 차세대 메모리 시장에서 trap layer 물질로 산화물 반도체를 사용하는 CTF의 연구 및 계발, 활용가치가 높을 것으로 기대된다.

• Journal of Korean Foot and Ankle Society
• /
• v.11 no.2
• /
• pp.182-186
• /
• 2007

Purpose: We examined the relationship of interdigital neuroma occurring site and the surrounding structures, including the deep transverse metatarsal ligament (DTML) by cadaver study and clinical results. Materials and Methods: Seventeen fresh frozen cadavers study were done to evaluate the relationship of interdigital neuroma occuring site and the DTML at two phase of the gait cycle with 60 degree of metatarsophalangeal dorsiflexion and with 15 degrees of ankle dorsiflexion. We measured the distance from interdigital nerve bifurcation of the common digital nerve to anterior margin of the DTML and longitudinal length of DTML itself. Clinically, we checked the location of interdigital neuroma and DTML length during surgery in 32 feet. Results: In the second and third web space, the mean distance from bifurcation of the common digital nerve of foot to the anterior margin of DTML was 16.7 mm, 15.1 mm in the mid-stance position, and 15.9 mm. 14.6 mm in heel-off position. Second, Third web space ligament itself length were average 12.8 mm, 10.6 mm. Clinically, all of the cases of interdigital neuroma started at the bifurcation area of the common digital nerve and interdigital neuroma was average 7.5 mm (range; 6-11 mm). Conclusion: Interdigital neuroma were located more distally than DTML in both the mid-stance and heel off stage. The main lesion was located between metatarsal head and metatarsophalangeal joint and more distal than the DTML anterior margin.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.43 no.1
• /
• pp.78-82
• /
• 2010

Vessel numbers and fuel consumption by South Korea's offshore and coastal fisheries have continuously declined since 2000. Using the 2006 Intergovernmental Panel on Climate Change Guidelines, $CO_2$ emissions by South Korea's fishery industry (fishing and aquaculture, excluding deep-sea fishing) were calculated by the default $ CO_2$ emission factor and fuel consumption by fuel type, Emission of $CO_2$ was estimated to be 3.22 million $tCO_2$/year in 2007 for fisheries (excluding deep-sea fishing); when including deep-sea fishing, the estimated value increased to 4.11 million $tCO_2$/year. Fuel consumption per tonne of fishery production was 498 L, and the amount of $CO_2$ emission per tonne of production was 1.62 $tCO_2$. To calculate $CO_2$ emission more exactly, we must develop a system to compile energy balance statistics and introduce life-cycle assessment for the fishery industry.

• The Bulletin of The Korean Astronomical Society
• /
• v.42 no.2
• /
• pp.61.1-61.1
• /
• 2017

As the application of deep-learning methods has been succeeded in various fields, they have a high potential to be applied to space weather forecasting. Convolutional neural network, one of deep learning methods, is specialized in image recognition. In this study, we apply the AlexNet architecture, which is a winner of Imagenet Large Scale Virtual Recognition Challenge (ILSVRC) 2012, to the forecast of daily solar flare occurrence using the MatConvNet software of MATLAB. Our input images are SOHO/MDI, EIT $195{\AA}$, and $304{\AA}$ from January 1996 to December 2010, and output ones are yes or no of flare occurrence. We consider other input images which consist of last two images and their difference image. We select training dataset from Jan 1996 to Dec 2000 and from Jan 2003 to Dec 2008. Testing dataset is chosen from Jan 2001 to Dec 2002 and from Jan 2009 to Dec 2010 in order to consider the solar cycle effect. In training dataset, we randomly select one fifth of training data for validation dataset to avoid the over-fitting problem. Our model successfully forecasts the flare occurrence with about 0.90 probability of detection (POD) for common flares (C-, M-, and X-class). While POD of major flares (M- and X-class) forecasting is 0.96, false alarm rate (FAR) also scores relatively high(0.60). We also present several statistical parameters such as critical success index (CSI) and true skill statistics (TSS). All statistical parameters do not strongly depend on the number of input data sets. Our model can immediately be applied to automatic forecasting service when image data are available.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.16 no.4
• /
• pp.479-490
• /
• 2018

The granite melting concept, which was suggested by Gibb's group for the closing of a deep borehole, was experimentally checked for KURT granite. The granite melting experiments were performed in two pressure conditions of atmospheric melting with certain inorganic additives and high pressure melting formed by water vaporization. The results of atmospheric tests showed that KURT granite started to melt at a lower temperature of $1,000^{\circ}C$ with NaOH addition and that needle shaped crystals were formed around partially melted crystals. In high pressure tests, vapor pressure was increased by adding water with maximum pressure of about 400 bars. KURT granite was partially melted at $1,000^{\circ}C$ when vapor pressure was low. However, it was not melted at vapor pressures higher than 200 bars. Therefore, it was determined that high pressure with a small amount of water vapor more effectively decreased the melting point of granite. Meanwhile, high temperature and high pressure vapor caused severe corrosion of the reactor wall.

• Journal of Korea Multimedia Society
• /
• v.25 no.1
• /
• pp.1-9
• /
• 2022

Ship detection in synthetic aperture radar (SAR) images is an important application in marine monitoring for the military and civilian domains. Over the past decade, object detection has achieved significant progress with the development of convolutional neural networks (CNNs) and lot of labeled databases. However, due to difficulty in collecting and labeling SAR images, it is still a challenging task to solve SAR ship detection CNNs. To overcome the problem, some methods have employed conventional data augmentation techniques such as flipping, cropping, and affine transformation, but it is insufficient to achieve robust performance to handle a wide variety of types of ships. In this paper, we present a novel and effective approach for deep SAR ship detection, that exploits label-rich Electro-Optical (EO) images. The proposed method consists of two components: a data augmentation network and a ship detection network. First, we train the data augmentation network based on conditional generative adversarial network (cGAN), which aims to generate additional SAR images from EO images. Since it is trained using unpaired EO and SAR images, we impose the cycle-consistency loss to preserve the structural information while translating the characteristics of the images. After training the data augmentation network, we leverage the augmented dataset constituted with real and translated SAR images to train the ship detection network. The experimental results include qualitative evaluation of the translated SAR images and the comparison of detection performance of the networks, trained with non-augmented and augmented dataset, which demonstrates the effectiveness of the proposed framework.