• KDI Journal of Economic Policy
• /
• v.40 no.2
• /
• pp.23-52
• /
• 2018

Given the limited availability of data in South Korea, this study proposes a method by which to estimate regional capital stock by modifying the benchmark year method (BYM) and applies it to estimate regional net capital stock by sector in transport infrastructure. First, it estimates time-varying sectoral depreciation rates using the sectoral net capital stock and the investment amount for each period. Second, it estimates the net capital stock of each period using the net capital stock in the base year and the investment in each period. Third, in order to ensure that the sum of net capital stocks by region is equal to the nationwide estimate, the national estimates are allocated to each region according to the proportion of the values derived from the previous stage. The proposed method can alleviate well-known problems associated with conventional BYMs, specifically the upward bias and arbitrary choice of the depreciation rate.

• IE interfaces
• /
• v.8 no.3
• /
• pp.61-73
• /
• 1995

As the level of maunfacturing system automation increases, the issues of modeling and simulation of AMS(Automated Manufacturing System) are becoming more important. Proposed in this paper is the JR-Net(Job Resource relation-Net) modeling framework which naturally mimics the process of designing an AMS by FA(Factory Automation) engineers. Its main purpose is to provide a modeling tool which facilitate modeling work of AMS for FA engineers unfamiliar with simulation modeling. The proposed modeling scheme is based on the extensive observation that typical AMSs are built from the set of 'standard' components(or catalog items). As an application of the proposed model, two real examples of FMS('G7'FMS model plant, RPI FMS) are modeled by JR-Net, and in case of FMS model plant, a simulation program development procedure using JR-Net modeling results is explained. Finally, simulation result of FMS model plant is analyzed.

• Smart Structures and Systems
• /
• v.16 no.1
• /
• pp.27-46
• /
• 2015

Active surface adjustment of cable net structures is becoming significant when large-size cable net structures are widely applied in various fields, especially in satellite antennas. A general-duty adjustment method based on active cables is proposed to achieve active surface adjustment or surface profile reconfiguration of cable net structures. Piezoelectric actuators and voice coil actuators are selected for constructing active cable structures and their simplified mechanical models are proposed. A bilevel optimization model of active surface adjustment is proposed based on the nonlinear static model established by the direct stiffness method. A pattern search algorithm combined with the trust region method is developed to solve this optimization problem. Numerical examples of a parabolic cable net reflector are analyzed and different distribution types of active cables are compared.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.272-275
• /
• 2018

본 논문에서는 기존 신경망 기반의 이미지 압축에 많이 사용되었던 신경망인 ResNet 을 대신하여 더 적은 개수의 파라미터를 사용하여 좋은 성능을 낼 수 있는 신경망 구조인 DenseNet 을 이미지 압축에 사용한다. 이미지 압축을 위해 사용되는 신경망 구조는 일반적으로 오토 인코더 구조인데, 병목 층에서 정보 손실이 상당히 많이 발생한다. 따라서 이미지 압축에서 신경망 내에서의 정보 전달은 상당히 중요하다. 기존의 논문에서는 이를 위해 이전의 정보를 그대로 뒤로 전달해주는 구조인 ResNet 을 사용하여 깊은 층에 대해서도 수렴이 잘 되는 결과를 보여주었다. 그러나 많은 수의 파라미터를 사용하는 단점을 해결하기 위해 본 논문에서는 DenseNet 을 이미지 압축에 사용하였고, 병목 층에서의 정보 손실로 인해 이미지의 고주파수 성분이 사라지는 현상을 해결하기 위해 원래 이미지와 JPEG2000 으로 압축한 이미지와의 차이를 추가 입력으로 넣어주어서 주관적인 화질을 개선하였다.

• International Journal of Computer Science & Network Security
• /
• v.21 no.2
• /
• pp.198-204
• /
• 2021

Glioma is one of the common types of brain tumors starting in the brain's glial cell. These tumors are classified into low-grade or high-grade tumors. Physicians analyze the stages of brain tumors and suggest treatment to the patient. The status of the tumor has an importance in the treatment. Nowadays, computerized systems are used to analyze and classify brain tumors. The accurate grading of the tumor makes sense in the treatment of brain tumors. This paper aims to develop a classification of low-grade glioma and high-grade glioma using a deep learning algorithm. This system utilizes four transfer learning algorithms, i.e., AlexNet, GoogLeNet, ResNet18, and ResNet50, for classification purposes. Among these algorithms, ResNet18 shows the highest classification accuracy of 97.19%.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.65-68
• /
• 2021

본 연구에서는 Dilated U-Net 기반의 이미지 복원기법을 통해 콘크리트 균열 추출 성능 개선 방안을 제안한다. 콘크리트 균열은 구조물의 미관상의 문제뿐 아니라 추후 큰 안전사고의 원인이 될 수 있어 초기대응이 중요하다. 현재는 점검자가 직접 육안으로 검사하는 외관 검사법이 주로 사용되고 있지만, 이는 정확성 및 비용, 시간, 그리고 안전성 면에서 한계를 갖고 있다. 이에 콘크리트 구조물 표면에 대해 획득한 영상 처리 기법을 사용한 검사 방식 도입의 관심이 늘어나고 있다. 또한, 딥러닝 기술의 발달로 딥러닝을 적용한 영상처리의 연구 역시 활발하게 진행되고 있다. 본 연구는 콘크리트 균열 추개선출 성능 개선을 위해 Dilated U-Net 기반의 이미지 복원기법을 적용하는 방안을 제안하였고 성능 검증 결과, 기존 U-Net 기반의 정확도가 98.78%, 조화평균 82.67%였던 것에 비해 정확도 99.199%, 조화평균 88.722%로 성능이 되었음을 확인하였다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.71-73
• /
• 2021

본 논문에서는 Scene Text Detection의 새로운 프레임워크인 TextFuseNet에 영상처리 관련 기술인 선명화 기법을 제안한다. Scene Text Detection은 야외 간판이나 표지판 등 불특정 배경에서 글자를 인식하는 기술이며, 그중 하나의 프레임워크가 TextFuseNet이다. TextFuseNet은 문자, 단어, 전역 기준으로 텍스트를 감지하는데, 여기서는 영상처리의 기술인 선명화 기법을 적용하여 TextFuseNet의 성능을 향상시키는 것이 목적이다. 선명화 기법은 기존 Sharpening Filter 방법과 Unsharp Masking 방법을 사용하였고 이 중 Sharpening Filter 방법을 적용하였을 때 AP가 0.9% 향상되었음을 확인하였다.

• ETRI Journal
• /
• v.45 no.4
• /
• pp.666-677
• /
• 2023

In the last few years, convolutional neural networks (CNNs) have demonstrated good performance while solving various computer vision problems. However, since CNNs exhibit high computational complexity, signal processing is performed on the server side. To reduce the computational complexity of CNNs for edge computing, a lightweight algorithm, such as a MobileNet, is proposed. Although MobileNet is lighter than other CNN models, it commonly achieves lower classification accuracy. Hence, to find a balance between complexity and accuracy, additional hyperparameters for adjusting the size of the model have recently been proposed. However, significantly increasing the number of parameters makes models dense and unsuitable for devices with limited computational resources. In this study, we propose a novel MobileNet architecture, in which the number of parameters is adaptively increased according to the importance of feature maps. We show that our proposed network achieves better classification accuracy with fewer parameters than the conventional MobileNet.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.31 no.4
• /
• pp.350-361
• /
• 1995

For the Purpose of the measurement of drifting condition of the anchovy drift gill net, some experiments were carried out in M. S Seong-Yang at the near sea of Kampo harbor in Korean east sea from August to October 1994, by using three set of GPS equipped with her net. The main results obtained are as follows : 1) In fixed position by GPS at Taebyon, the radius of 95% probability and shifted distance from standard position were respectively 79.8m, 21.0m. 2) The mean values of PDOP and HDOP of GPS at Taebyon were respectively 3.5, 1.9 and the stability of the position by GPS at Taebyon was found to be very high. 3) In the fishing ground of the position from Lat. $35^{\circ}$ 41.5'N to $Lat^{\circ}$. 43.5', from Long. $129^{\circ}$ 32.0'E to Long. $129^{\circ}$ 32.5'E, the direction of the whole stretched drift gill net was $190^{\circ}$ and the direction of the whole casted net was $200^{\circ}$ when casting the net was finished. When whole net was found to be drifting into the direction of $170^{\circ}\;to\;180^{\circ}$, and the mean drifting speed of that was 0.82kt. 4) In the fishing ground of the position from Lat. $35^{\circ}$ 44.0'N to Lat. $35^{\circ}$ 45.2'N, from Long. $129^{\circ}$ 30.1'E to Long. $129^{\circ}$ 31.2'E, the direction of the whole stretched net was 20$^{\circ}$ and the direction of the whole stretched net was $20^{\circ}$ when casting the net was finished. When hauling the net, the direction of the spreaded net appeared from the North into the South and the whole net was found to be drifting into the direction of $210^{\circ}\;to\;220^{\circ}$, and the mean drifting speed of that was 0.75kt. 5) In the fishing ground of the positioning from Lat. $35^{\circ}$ 44.9'N to Lat. $35^{\circ}$ 46.0'N, Long. $129^{\circ}$ 33.0'E to Long. $129^{\circ}$ 34.0'E, the direction to the whole stretched net was $35^{\circ}$ and the direction of the whole casted net was $50^{\circ}$ when casting the net was finished. When hauling the net, the direction of the spreaded net appeared into the South from the East and the whole net was found to be drifting into the direction of $330^{\circ}\;to\;40^{\circ}$, and the mean drifting speed of the was 0.63kt.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.60 no.4
• /
• pp.504-509
• /
• 2015

This research carried out to figure out the effect of the green manure crop cultivated at a preparation field and the shading net on the growth, development, and quality of ginseng. Followings are results obtained from the research. Leaf width of ginseng under the shading net of a two-layered blue and two-layered black polythylene net (TBTBPN) was good at rye and hairy vetch cultured group. Leaf length of ginseng under the shading net of a three-layered blue and one-layered black polyethylene net (TBOBPN) was good at barley and hairy vetch cultured group. Meanwhile, leaf width was good at hairy vetch cultured group. Leaf length of ginseng under a blue polyethylene sheet (BPS) was good at a barley and barley + hairy vetch cultured group, but stem length was shorter compare to other shading net cultivations. Root weight of ginseng was good under the shading net of a two-layered blue and two-layered black polyethylene net (TBTBPN) at a rye and hairy vetch cultured group, and was good under the shading net of a three-layered blue and onelayered black polyethylene net (TBOBPN) at a barley + hairy vetch cultured group, but there was no significant difference under blackout screen according to manure crop varieties. Ratio of rusty root was 10.2% at the barley cultured group under the shading net of a two-layered blue and two-layered black polyethylene net (TBTBPN), and was 23.1% at hairy vetch cultured group under shading net of a three-layered blue and one-layered black polyethylene net (TBOBPN). Ratio of rusty root was the lowest at a rye cultured group regardless the shading nets. Content of the ginsenoside was the highest at the rye cultured group under the shading net of two-layered blue and two-layered black polyethylene net (TBTBPN), was the highest at the barley cultured group under the shading net of a three-layered blue and one-layered black polyethylene net (TBOBPN), and was the highest at the rye cultured group under the blackout screen.