• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권8호
• /
• pp.3984-4005
• /
• 2017

Image recognition has become an increasingly important topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The recognition systems rely on a key component, i.e. the low-level feature or the learned mid-level feature. The recognition performance can be potentially improved if the data distribution information is exploited using a more sophisticated way, which usually a function over hidden variable, model parameter and observed data. These methods are called generative score space. In this paper, we propose a discriminative extension for the existing generative score space methods, which exploits class label when deriving score functions for image recognition task. Specifically, we first extend the regular generative models to class conditional models over both observed variable and class label. Then, we derive the mid-level feature mapping from the extended models. At last, the derived feature mapping is embedded into a discriminative classifier for image recognition. The advantages of our proposed approach are two folds. First, the resulted methods take simple and intuitive forms which are weighted versions of existing methods, benefitting from the Bayesian inference of class label. Second, the probabilistic generative modeling allows us to exploit hidden information and is well adapt to data distribution. To validate the effectiveness of the proposed method, we cooperate our discriminative extension with three generative models for image recognition task. The experimental results validate the effectiveness of our proposed approach.

• 한국항공우주학회지
• /
• 제38권6호
• /
• pp.620-628
• /
• 2010

소형위성은 저비용으로 단기간에 개발이 가능하기 때문에 이를 통해 폭 넓은 우주 연구를 수행할 수 있고 다양한 분야의 활용 가능하다. 소형위성은 빠른 개발이 가능한 이점이 있다. 따라서 신속 대응 우주 구현이 주목됨에 따라 소형위성의 역할이 급속히 증가하고 있다. 하지만 소형위성은 기존의 위성에 비해 크기나 전력의 제약이 많기 때문에 고성능의 안정적인 시스템을 기대하기는 어렵다. 이러한 문제점을 해결하고 빠른 개발을 위해 표준화와 모듈화 설계가 필요하다. 모듈화는 플러그 앤 플레이(Plug-and-Play)를 지원하여 수일 내에 인공위성 제작 및 시험을 가능하게 한다. 이러한 모듈 간의 원활한 데이터 통신을 위해서는 데이터 버스의 표준이 요구된다. CAN 통신 방식은 플러그 앤 플레이에 가장 효과적으로 대처 할 수 있는 통신방식으로 꼽힌다. CAN은 높은 신뢰성을 가지며 분산 시스템을 지원하여 위성의 호환성을 높여준다. 따라서 시험이 용이해지고 짧은 기간에 고성능의 안정적인 위성 개발이 가능하게 된다. 본 논문에서는 모듈화 방식의 소형위성 개발 방안에 대해 제안하고, CAN을 데이터 버스로 적용하여 소형위성 내부 데이터 인터페이스 버스를 설계하고 시험을 통해 적합성을 분석하여 기술하였다.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2021년도 하계학술대회
• /
• pp.23-25
• /
• 2021

This paper describes an approach of composing music with multiple AI composers. This approach enriches more the creativity space of artificial intelligence music composition than using only one composer. This paper presents a simple example with 2 different deep learning composers working together for composing one music. For the experiment, the two composers adopt the same deep learning architecture of an LSTM model trained with different data. The output of a composer is a sequence of notes. Each composer alternatively appends its output to the resulting music which is input to both the composers. Experiments compare different music generated by the proposed multiple composer approach with the traditional one composer approach.

• 한국실내디자인학회논문집
• /
• 제15권4호
• /
• pp.97-105
• /
• 2006

Richard Meier has earned his recognition by developing his unique design approach and keeping consistency in design concept in museum architecture over the years of his practice. The purpose of this study is finding common characteristics of the exhibition space spatial organization based on actual site visit analysis. This study will mainly focus on the museums Richard Meier designed between 1979, which is the year he started, until 1997 which is the year Getty Museum opened. For the purpose of this study, Museum of Applied Art, Frankfurt(1979-85), High Museum of Art, Atlanta(1980-83), Museum of Contemporary Art, Barcelona(1987-95), and The J.Paul Getty Museum(1984-1997) have been selected for analysis. This study will reveal fundamental characteristics of Richard Meier's architectural philosophy through his early residential projects and give background information of the 'white architecture' Close analysis will continue by reviewing four museums selected specifically for this study and focus more from the point of issues such as; 1)applying various axis in space programming, 2)symbolic significance of exhibition space, 3)hierarchy of each space, 4)design elements and characteristics of the exhibition space. Furthermore, quantitative research on exhibition space organization will be done by Space Syntax theory method. The result of the study will be a significant data in designing a modern museum.

• KIEAE Journal
• /
• 제12권5호
• /
• pp.29-34
• /
• 2012

A rental housing has a key role to supply the living space to non-ownership. A public rental housing is particularly aimed at providing the living space for the low-income households with a low rent fee. Therefore, the local government would try to maintain the facilities of the rental housing and to get the decent living condition. For getting the required living condition, it should make a maintenance plan, which contains the repair time, repair scope and targeted component and finishings. This study is aimed at providing the optimal repair time in 12 components and components of the rental housing which is controlled by the local government. The optimal repair time has two steps to get the final result. First, it would draw the 1'st repair time with the probabilistic and empirical approach. Second, comparing the drawn data and the service life, the optima repair time would be provided with considering the components' attributes. Result are as follows : First, the optimal repair time would be considered with the component attributes. There are user's convenience, its safety and physical aspect. Second, the kitchen utensils, elevator and water tap has a optimal time of 16, 19, 17 years respectively which is considered with physical aspect. In addition, the optima repair time of the wiring appliance and lighting equipment are 12 years and 10 years respectively.

• ETRI Journal
• /
• 제33권5호
• /
• pp.741-751
• /
• 2011

This paper focuses on online speaker tracking for telephone conversations and broadcast news. Since the online applicability imposes some limitations on the tracking strategy, such as data insufficiency, a reliable approach should be applied to compensate for this shortage. In this framework, a set of reference speaker models are used as side information to facilitate online tracking. To improve the indexing accuracy, adaptation approaches in eigenvoice decomposition space are proposed in this paper. We believe that the eigenvoice adaptation techniques would help to embed the speaker space in the models and hence enrich the generality of the selected speaker models. Also, an index structure of the reference models is proposed to speed up the search in the model space. The proposed framework is evaluated on 2002 Rich Transcription Broadcast News and Conversational Telephone Speech corpus as well as a synthetic dataset. The indexing errors of the proposed framework on telephone conversations, broadcast news, and synthetic dataset are 8.77%, 9.36%, and 12.4%, respectively. Using the index tree structure approach, the run time of the proposed framework is improved by 22%.

• 천문학회보
• /
• 제40권2호
• /
• pp.57.1-57.1
• /
• 2015

VLBI에서 관측 대역폭이 n배 증가될 경우 관측감도는 sqrt(n)만큼 향상되고, 이는 관측 연구 측면에서 기존에는 불가능하던 천체에 대한 연구 수행이 가능함을 의미한다. 관측 대역폭의 확대는 관측 데이터의 용량 증가를 의미하며 여기서 해당 데이터의 처리를 위한 초고속 데이터 기록시스템은 핵심적 역할을 한다. 이에 따라 현재 KVN에서는 미국 MIT Haystack 천문대에서 개발된 초고속 기록시스템인 Mark5B/B+와 Mark6를 운용 중에 있다. 하지만 이들 시스템의 경우 사실상 VLBI연구를 위한 목표로 특수 개발되었기에 유지 및 운영 측면에서 여러 불편이 있고, 성능에 있어서도 단일 스트림 기준으로 8Gbps를 넘지 못하는 한계를 안고 있다. 본 발표에서는 기존 시스템을 대체할 수 있는 기술로서 PCI 익스프레스 기반의 데이터 처리를 소개하고자 한다. 나아가 실제 관측 데이터에 대한 입출력 및 기존 시스템과의 성능 비교를 통해 광대역 관측 연구에 최적화된 기록 시스템을 제안하고자 한다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.29-31
• /
• 2007

In this paper we propose an directional analytic approach in image data space for X-ray image which is detected from the X-ray projection system. Such a radiographic nondestructive testing has long been used for steel-tube inspection and weld monitoring. The welded area and thickness of steel-tube are detected from gradient radiating mechanism based on the evaluation of biased X-ray source position. The welded area is an ellipse type on low contrast X-ray image including noise. Noise originates from most of elements of the system. such as shielding CCD camera, imaging screen, X-ray source, inspected object, electronic circuits and etc.. Projection incorrectness and noise influence on imaging quality is to be represented by vertical pixels' distribution. Space analysis due to vertical direction also shows the segmental possibility between regions by visual edge evaluation.

• Journal of Positioning, Navigation, and Timing
• /
• 제10권2호
• /
• pp.83-89
• /
• 2021

The extended Kalman filter (EKF) is widely used for global navigation satellite system (GNSS) applications. It is difficult to obtain precise positions with an EKF one-way (forward or backward) filter. In this paper, we propose an EKF smoother to improve the positioning accuracy by integrating forward and backward filters. For the EKF smoother experiment, we performed PPP using GNSS data received at the DAEJ reference station for a month. The effectiveness of the proposed approach is validated with multi-GNSS kinematic PPP experiments. The EKF smoother showed 35%, 6%, and 22% improvement in east, north, and up directions, respectively. In addition, accurate tropospheric zenith total delay (ZTD) values were calculated by a smoother. Therefore, the results from EKF smoother demonstrate that better accuracy of position can be achieved.

• 항공우주시스템공학회지
• /
• 제16권4호
• /
• pp.17-27
• /
• 2022

본 논문에서는 에어포일의 좌표 데이터에 대해 공력 특성을 예측할 수 있는 합성곱 신경망 기반 네트워크 프레임 워크를 설계하였으며 Xfoil을 이용한 공력 데이터를 적용하여 네트워크의 가능성을 확인하였다. 이 때 에어포일의 두께 변화에 따른 공력 특성 예측을 수행하였다. 부호화 거리 함수를 이용하여 에어포일의 좌표 데이터를 이미지 데이터로 변환하였으며 받음각 정보를 반영하였다. 또한 에어포일의 압력 계수 분포를 축소 모델 기법 중 하나인 적합 직교 분해를 이용하여 축소된 데이터로 표현하였으며 이를 네트워크의 출력 데이터로 사용하였다. 제시하는 네트워크의 내삽과 외삽 성능을 평가하기 위하여 시험 데이터를 구성하였고, 결과적으로 내삽 데이터에 대한 예측 성능이 외삽에 비해 우수함을 확인하였다.