• 한국전자파학회지:전자파기술
• /
• 제6권3호
• /
• pp.80-90
• /
• 1995

마이크로스트립 단일선로, 직각벤드 및 결합선로의 산란행렬을 multipart r$\xi$twork model을 이용하여 1-18GHz 범위에서 계산했다. 단일선로는 평면형 도파관 모텔을 이용하여 해석했다. 직각벤드의 경우는 두개 의 사각형 부분으로 나누어져 각 부분은 단일선로와 같은 방법으로 해석되며, 각 부분의 임피던스 행렬은segmentation 방볍을 이용하여 연결된다. 결합선로 사이의 전자계 결합의 해석에는 Green 함수를 이용하는 기 존의 방법보다 계산시간이 적게 소요되는 새로운 방법이 이용되었다 상기의 세가지 구조에 대한 수치해석 결 과는 SuperCompact의 결과와 잘 일치하며, 이는 상기의 방법이 복잡한 단일 및 결합선로 불연속구조의 해석에 유용하게 사용될 수 있음을 보여준다

• 한국정보처리학회논문지
• /
• 제4권1호
• /
• pp.280-287
• /
• 1997

성공적인 음성인식·합성 시스템을 구축하기 위해서는 음운론적 지식, 특히 운율 정보의 도입이 매우 중요하다. 본 연구에서는 우선 음성인식·합성을 위한 운율음운 론의 연구동향을 개관하고, 국어의 음운단위와 경계의 설정에 관한 이론적·실험적 고찰을 정리하였으며, 음운단위에 따른 경계의 자동적 발견을 위하여, 데이터를 수집 하고 시스템을 구현하여 실험을 행하였다. 단순회귀 신경망을 이용하여, 2,200여 개 의 문장에 있는 12,000여개의 음운단어를 외부정보의 도움이 전혀 없이 훈련시킨 결 과, 70%정도의 예측률을 보였다. 본 연구에서 사용한 방법을 다른 정보와 결합하여 사용한다면, 음운경계의 발전과 그에 따른 분절화를 정확하게 행할 수 있으리라 기대 된다.

• 한국농업기계학회:학술대회논문집
• /
• 한국농업기계학회 1996년도 International Conference on Agricultural Machinery Engineering Proceedings
• /
• pp.773-779
• /
• 1996

Dried oak mushroom have complex and various visual features. Grading and sorting of dried oak mushrooms has been done by the human expert. Though actions involved in human grading looked simple, a decision making underneath the simple action comes from the result of the complex neural processing of the visual image. Through processing details involved in human visual recognition has not been fully investigated yet, it might say human can recognize objects via one of three ways such as extracting specific features or just image itself without extracting those features or in a combined manner. In most cases, extracting some special quantitative features from the camera image requires complex algorithms and processing of the gray level image requires the heavy computing load. This fact can be worse especially in dealing with nonuniform, irregular and fuzzy shaped agricultural products, resulting in poor performance because of the sensitiveness to the crisp criteria or specific ules set up by algorithms. Also restriction of the real time processing often forces to use binary segmentation but in that case some important information of the object can be lost. In this paper, the neuro net based real time recognition algorithm was proposed without extracting any visual feature but using only the directly captured raw gray images. Specially formated adaptable size of grids was proposed for the network input. The compensation of illumination was also done to accomodate the variable lighting environment. The proposed grading scheme showed very successful results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권6호
• /
• pp.2018-2043
• /
• 2022

Nowadays, COVID-19 infections are influencing our daily lives which have spread globally. The major symptoms' of COVID-19 are dry cough, sore throat, and fever which in turn to critical complications like multi organs failure, acute respiratory distress syndrome, etc. Therefore, to hinder the spread of COVID-19, a Computerized Doughty Predictor Framework (CDPF) is developed to yield benefits in monitoring the progression of disease from Chest CT images which will reduce the mortality rates significantly. The proposed framework CDPF employs Convolutional Neural Network (CNN) as a feature extractor to extract the features from CT images. Subsequently, the extracted features are fed into the Adaptive Dragonfly Algorithm (ADA) to extract the most significant features which will smoothly drive the diagnosing of the COVID and Non-COVID cases with the support of Doughty Learners (DL). This paper uses the publicly available SARS-CoV-2 and Github COVID CT dataset which contains 2482 and 812 CT images with two class labels COVID⁺ and COVI^-. The performance of CDPF is evaluated against existing state of art approaches, which shows the superiority of CDPF with the diagnosis accuracy of about 99.76%.

• 한국멀티미디어학회논문지
• /
• 제22권5호
• /
• pp.535-546
• /
• 2019

Recently, there have been a lot of researches on the whole face replacement system, but it is not easy to obtain stable results due to various attitudes, angles and facial diversity. To produce a natural synthesis result when replacing the face shown in the video image, technologies such as face area detection, feature extraction, face alignment, face area segmentation, 3D attitude adjustment and facial transposition should all operate at a precise level. And each technology must be able to be interdependently combined. The results of our analysis show that the difficulty of implementing the technology and contribution to the system in facial replacement technology has increased in facial feature point extraction and facial alignment technology. On the other hand, the difficulty of the facial transposition technique and the three-dimensional posture adjustment technique were low, but showed the need for development. In this paper, we propose four facial replacement models such as 2-D Faceswap, OpenPose, Deekfake, and Cycle GAN, which are suitable for the Cox platform. These models have the following features; i.e. these models include a suitable model for front face pose image conversion, face pose image with active body movement, and face movement with right and left side by 15 degrees, Generative Adversarial Network.

• Korean Journal of Radiology
• /
• 제24권11호
• /
• pp.1131-1141
• /
• 2023

Objective: Cortical iron deposition has recently been shown to occur in Alzheimer's disease (AD). In this study, we aimed to evaluate how cortical gray matter iron, measured using quantitative susceptibility mapping (QSM), differs in the clinical cognitive impairment spectrum. Materials and Methods: This retrospective study evaluated 73 participants (mean age ± standard deviation, 66.7 ± 7.6 years; 52 females and 21 males) with normal cognition (NC), 158 patients with mild cognitive impairment (MCI), and 48 patients with AD dementia. The participants underwent brain magnetic resonance imaging using a three-dimensional multi-dynamic multi-echo sequence on a 3-T scanner. We employed a deep neural network (QSMnet+) and used automatic segmentation software based on FreeSurfer v6.0 to extract anatomical labels and volumes of interest in the cortex. We used analysis of covariance to investigate the differences in susceptibility among the clinical diagnostic groups in each brain region. Multivariable linear regression analysis was performed to study the association between susceptibility values and cognitive scores including the Mini-Mental State Examination (MMSE). Results: Among the three groups, the frontal (P < 0.001), temporal (P = 0.004), parietal (P = 0.001), occipital (P < 0.001), and cingulate cortices (P < 0.001) showed a higher mean susceptibility in patients with MCI and AD than in NC subjects. In the combined MCI and AD group, the mean susceptibility in the cingulate cortex (β = -216.21, P = 0.019) and insular cortex (β = -276.65, P = 0.001) were significant independent predictors of MMSE scores after correcting for age, sex, education, regional volume, and APOE4 carrier status. Conclusion: Iron deposition in the cortex, as measured by QSMnet+, was higher in patients with AD and MCI than in NC participants. Iron deposition in the cingulate and insular cortices may be an early imaging marker of cognitive impairment related neurodegeneration.

• 한국수자원학회논문집
• /
• 제53권7호
• /
• pp.545-555
• /
• 2020

본 연구에서는 해외 수력댐 건설 프로젝트의 사전조사 기초자료 제공을 위하여 댐 위치 결정을 위한 사전적지분석 알고리즘을 개발하고, 위성영상 수치표고자료인 ASTER Global Digital Elevation Model (ASTGTM)과 토지피복자료인 Terra/Aqua combined Moderate Resolution Imaging Spectroradiometer (MODIS) MCD12Q1를 사용하였다. 사전적지분석 알고리즘은 DEM의 전처리, 하천망생성, 유역분할과 지형정보를 고려한 적지분석과 댐 건설 시 수몰면적에 따른 보상면적 산정 알고리즘을 포함하고 있으며 Python기반의 오픈소스 GIS로 구현되었다. 적지분석은 사용자가 하천 위의 지점을 선택하면, DEM으로부터 낙차, 도달시간, 내용적곡선과 같은 지형정보와 토지피복자료를 통한 보상면적을 기반으로 지점의 적지여부를 평가한다. 분석알고리즘은 국내 부항, 보현산, 성덕, 영주댐을 대상으로 시범적용 됐으며 해당 지점의 가능 최대낙차는 각각 37, 67, 73, 42 m로 나타났으며 최대저수면적은 1.81, 2.4, 2.8, 8.8 ㎢ 최대저수량은 35.9, 68, 91.3, 168.3×10⁶ ㎥으로 나타났다. 보현산과 성주 댐에서는 타당성을 보였으나, 부항과 영주 댐의 경우 ASTGTM 에러로 인한 잘못된 하천망과 유역경계로 인해 낙차가 제한됨을 보였다, 본 연구의 결과는 향후 해외 수력댐 사업 진출시 사전분석에서 적지의 지형학적 평가에 도움이 될 것으로 기대된다.