• Journal of the Korea Society of Computer and Information
• /
• v.12 no.3
• /
• pp.277-285
• /
• 2007

This research proposes a personalized mobile web contents creation method that uses web news channel contents-based analysis. It promptly acquires data through the RSS and RSS-linked web pages which have been supplied by the existing web sites for a news search. And then It applies a personalization method using analysis in contents filtering and generation. The proposed method will make creating mobile web contents easier while lowering wireless contents production costs. Moreover, It can be improved a user satisfaction for contents filtering and access with using analysis that fits in with a matter of user's specific interest.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.1 no.3
• /
• pp.53-60
• /
• 2008

Recently, the application of internet services has been realized by effort for an offer of various informations and personalization services. So in this paper, we propose the web service application model for service integration and personalization based on internet multimedia subsystems. For this services integration and personalization, we need to establish a policy of supporting personalization service. And It's required to analyze information of service users and to extract the components for the personalization services. With the process, It will provide the efficient integration between the exist services and the external services and It can be realized detail services for personal with construction of new services model.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.41 no.10
• /
• pp.63-68
• /
• 2004

A high-resolution synchronous mirror delay (SMD) is proposed in order to reduce the clock skew between the external clock and the infernal clock of a chip. The proposed SMD reduces the clock skew in two steps. Coarse locking is achieved by the SMD. Fine locking is achieved by the successive approximation register-controlled DLL. The total locking time is 10 clock cycles. Simulation results show that the proposed SMD operates with 50psec clock skew at 182MHz and consumes 17.5mW at 3.3V supply voltage in a 0.35 um 1-poly 4-metal CMOS technology.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.51-54
• /
• 2024

본 논문에서는 온 디바이스 국방 AI를 위한 효율적인 학습 방법을 제안한다. 제안하는 방법은 모델 전체를 재학습하는 대신 필요한 부분만 세밀하게 조정하여 계산 비용과 시간을 대폭 줄이는 PEFT 기법의 LoRa를 적용하였다. LoRa는 기존의 신경망 가중치를 직접 수정하지 않고 추가적인 낮은 랭크의 매트릭스를 학습하는 방식으로 기존 모델의 구조를 크게 변경하지 않으면서도, 효율적으로 새로운 작업에 적응할 수 있다. 또한 학습 파라미터 및 연산 입출력에 데이터에 대하여 32비트의 부동소수점(FP32) 대신 부동소수점(FP16, FP8) 또는 정수형(INT8)을 활용하는 경량화 기법인 양자화도 적용하였다. 적용 결과 학습시 요구되는 GPU의 사용량이 32GB에서 5.7GB로 82.19% 감소함을 확인하였다. 동일한 조건에서 동일한 데이터로 모델의 성능을 평가한 결과 동일 학습 횟수에선 LoRa와 양자화가 적용된 모델의 오류가 기본 모델보다 53.34% 증가함을 확인하였다. 모델 성능의 감소를 줄이기 위해서는 학습 횟수를 더 증가시킨 결과 오류 증가율이 29.29%로 동일 학습 횟수보다 더 줄어듬을 확인하였다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.34 no.12
• /
• pp.641-655
• /
• 2007

Parameterization has been one of very important research subjects in several application areas including computer graphics. In the parameterization research, the problem of mapping 3D polygonal model to 2D plane has been studied frequently, but the previous methods often fail to handle complicated shapes of polygonal surfaces effectively as well as entail distortion between the 3D and 2D spaces. Several attempts have been made especially to reduce such distortion, but they often suffer from the problem when an arbitrary rectangular surface region on 3D model is locally parameterized. In this paper, we propose a new local parameterization scheme based on the projection level set method. This technique generates a series of equi-distanced curves on the surface region of interest, which are then used to generate effective local parameterization information. In this paper, we explain the new technique in detail and show its effectiveness by demonstrating experimental results.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.137-137
• /
• 2017

현재 상용화되어 있는 온실 환경제어시스템의 S/W 및 H/W는 서로 호환이 되지 않아 농민들이 원하는 맞춤형 복합 환경 제어시스템을 운영하는데 어려움이 있다. 따라서 본 연구는 다양한 제어알고리즘 및 장비를 적용시킬 수 있는 호환성이 향상된 온실 환경 제어인터페이스 모듈 성능평가를 목표로 한다. 센서 및 제어 인터페이스 모듈 성능평가를 위해 사용된 제어 시스템은 8 bit MCU가 적용된 전용 개발보드를 사용하였고, RS-232 통신 케이블을 사용하여 온실 환경 측정 데이터 값을 PC에서 수신할 수 있도록 하였다. 또한, 창개폐기, 환풍기를 사용하여 온실 내부 온/습도 환경조성을 하였다. 실험은 정오부터 제어장비를 작동시킨 후 1시간 간격으로 총 3시간 동안의 온실 내 온/습도의 변화량을 계측하였다. 3시간 중 1시간동안의 온/습도 값의 변화량을 계측한 결과 평균값은 각각 $33.21^{\circ}C$, 34.94%이었고 표준편차는 각각 $1.44^{\circ}C$, 2.74% 이었다. 제어 알고리즘은 단순한 ON/OFF 방법을 사용 하였고 총 2가지 제어장비를 사용하였으며 모두 정상 작동 하였다. 1시간동안 온실의 온도는 $30^{\circ}C{\sim}35^{\circ}C$사이를 유지하였으며, 습도는 30%~ 40% 사이로 최초 실험 목표였던 온실 내부의 온/습도 범위를 유지하였다. 이번 실험은 ON/OFF 방법의 제어 알고리즘을 사용하였지만 더욱 정밀한 온실 환경제어를 위하여 PID, 퍼지 제어 알고리즘을 추가하여 기상환경에 따른 제어범위를 더욱 세밀화 할 수 있도록 설계한다면 제어장비에 대한 효율성이 향상될 것이라 기대한다.

• Applied Biological Chemistry
• /
• v.34 no.4
• /
• pp.312-317
• /
• 1991

The volatile components of Korean and Chinese Inulae flos were identified. It is made of dried Inulae flos from the components were collected by simultaneous steam distillation extract method. Those were analyzed by combined gas chromatography(GC) and gas chromatography-mass spectrometrt(GC/MS). Sixty two components, including 10 hydrocarbons, 7 aldehydes, 18 alcohols, 3 ketones, 5 esters, 13 acids and 6 miscellaneous components were identified. Volatile components in Inulae flos were fractinated into one hydrocarbon fraction and two oxygenated hydrocarbon fractions by using silica gel column chromatography. The volatile components consisted of 16.85% hydrocarbon and 83.15% oxygenated hydrocarbons in Korean Inulae flos, 23.46% hydrocarbon and 76.53% oxygenated hydrocarbons in Chinese Inulae flos.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.23 no.3
• /
• pp.146-160
• /
• 2020

In the 4th industrial revolution, research on solving urban problems using ICT technology is emerging. Representatively, research is being conducted mainly focusing on smart cities. 3D geo-spatial. Various case studies on smart cities are being conducted to solve urban problems, and in order to effectively implement smart cities, it is implemented on the basis of a digital twin interconnecting physical and virtual environments. To implement a digital twin, the concept of space is essential, and spatial information that composes the real world is also essential. Therefore, in order to implement the digital twin, this study conducted a study on the construction and modeling of 3D geo-spatial information based on the national spatial information establishment regulations and the international standard CityGML for buildings, water facilities, and transportation facilities in Jeonju. In Jeonju, the research area, the LOD was defined, and the possibility of digital twin based on geo-spatial information was determined through data collection, analysis, and construction for each detail. In this study, it is meaningful that a plan to construct geo-spatial information for the implementation of a digital twin is presented, and the result of the construction is modeled as an urban standard model. The results of study are expected to be used as reference for detailed construction plans and standard establishment by governments and local governments promoting smart city and digital twin construction in the future.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.10
• /
• pp.520-529
• /
• 2005

A feature-based 3D modeling algorithm is presented in this paper. Since conventional methods use depth-based techniques, they need much time for the image matching to extract depth information. Even feature-based methods have less computation load than that of depth-based ones, the calculation of modeling error about whole pixels within a triangle is needed in feature-based algorithms. It also increase the computation time. Therefore, the proposed algorithm consists of three phases, which are an initial 3D model generation, model evaluation, and model refinement phases, in order to acquire an efficient 3D model. Intensity gradients and incremental Delaunay triangulation are used in the Initial model generation. In this phase, a morphological edge operator is adopted for a fast edge filtering, and the incremental Delaunay triangulation is modified to decrease the computation time by avoiding the calculation errors of whole pixels and selecting a vertex at the near of the centroid within the previous triangle. After the model generation, sparse vertices are matched, then the faces are evaluated with the size, approximation error, and disparity fluctuation of the face in evaluation stage. Thereafter, the faces which have a large error are selectively refined into smaller faces. Experimental results showed that the proposed algorithm could acquire an adaptive model with less modeling errors for both smooth and abrupt areas and could remarkably reduce the model acquisition time.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.3
• /
• pp.239-244
• /
• 2014

As the usefulness of mobile device is kept growing the privacy of the cloud computing is receiving more attentions. Even though many researches and solutions for privacy matters are suggested we are still worrying about the security problems. In addition most of cloud computing systems uses file-level synchronization which make it difficult to modify a part of a file. If we use data-centric app that stores data on embedded DBMS such as SQLite, a simple synchronization may incur some loss of information. In this paper we propose a solution to build a personal cloud that supports record-level synchronization. And we show a prototype system which uses RESTful web services and the same schema on mobie devices and the cloud storage. Synchronization is achieved by using a kind of optimistic concurrency control.