• Journal of the Korea Society of Computer and Information
• /
• v.22 no.3
• /
• pp.45-51
• /
• 2017

In this paper, we propose a consecutive-frame super-resolution method to tackle a moving object problem. The super-resolution is a method restoring a high resolution image from a low resolution image. The super-resolution is classified into two types, briefly, single-frame super-resolution and consecutive-frame super-resolution. Typically, the consecutive-frame super-resolution recovers a better than the single-frame super-resolution, because it use more information from consecutive frames. However, the consecutive-frame super-resolution failed to recover the moving object. Therefore, we proposed an improved method via moving object detection. Experimental results showed that the proposed method restored both the moving object and the background properly.

• Journal of Digital Convergence
• /
• v.11 no.4
• /
• pp.215-220
• /
• 2013

The super computer calls usually as the super computer in case the computing power of the computer is 20 G flops (GFLOPS) or greater. In the past, the computer equipped with the vector processor (the instrument processing the order having the logic operation and maximum value or minimum value besides the common computer instruction) processing the scientific calculation with the super high speed was installed as the super computer. Recently, cyber attack focuses on supercomputer because if it is being infected, then it will affect hundreds of client PC. Therefore, our research paper analyzed super computer security issues and biometric countermeasure to develop the level of security on super computer.

• Journal of Sensor Science and Technology
• /
• v.19 no.2
• /
• pp.118-123
• /
• 2010

This paper proposes super-resolution reconstruction algorithm for image enhancement. Super-resolution reconstruction algorithms reconstruct a high-resolution image from multi-frame low-resolution images of a scene. Conventional super- resolution reconstruction algorithms are iterative back-projection(IBP), robust super-resolution(RS)method and standard Papoulis-Gerchberg(PG)method. However, traditional methods have some problems such as rotation and ringing. So, this paper proposes modified algorithm to improve the problem. Experimental results show that this proposed algorithm solve the problem. As a result, the proposed method showed an increase in the PSNR for traditional super-resolution reconstruction algorithms.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.11a
• /
• pp.932-935
• /
• 2007

모바일 기기 사용이 급증함에 따라 모바일환경에서 이루어지는 P2P방식 연구가 활발히 진행되고 있다. 본 논문에서는 기존의 모바일환경에서의 P2P방식이 지닌 peer들 사이의 multi-broadcasting 방식의 문제점을 보완하고, 새로운 routing table을 구축하기 위해 peer들을 2개의 계층으로 구분하였다. 즉, peer들을 super peer들과 각 super peer에 의해 관리되는 sub peer들로 구분하였다. 파일의 탐색과 전송은 super peer들이 관리하므로, 기존의 불필요한 multi-broadcasting 방식을 피할 수 있다. 본 논문에서는 분산 네트워크를 위한 peer의 계층화 작업을 설계하였다. MIS(Maximal Independent Set)알고리즘을 이용하여 peer들의 계층화 작업을 외부의 도움 없이 peer들간의 통신으로 할 수 있게 만들었다. 이처럼 peer들을 super peer와 sub peer로 구분하면 불필요한 broadcasting을 피할 수 있어 시스템 성능이 향상되며, 이를 실험을 통하여 증명하였다.

• International Journal of High-Rise Buildings
• /
• v.4 no.1
• /
• pp.75-83
• /
• 2015

In the present study, the preliminary results of a large-scale seismic response analysis of a super-high-rise steel frame considering soil-structure interaction are presented. A seismic response analysis under the excitation of the JR Takatori record of the 1995 Hyogoken-Nanbu earthquake is conducted. Precise meshes of a 31-story super-high-rise steel frame and a soil region, which are constructed completely of hexahedral elements, are generated and combined. The parallel large-scale simulation is performed using K computer, which is one of the fastest supercomputers in the world. The results are visualized using an offline rendering code implemented on K computer, and the feasibility of using a very fine mesh of solid elements is investigated. The computation performance of the analysis code on K computer is also presented.

• Journal of Information Processing Systems
• /
• v.14 no.5
• /
• pp.1114-1135
• /
• 2018

A multiple classification system based on a new boosting technique has been approached utilizing different biometric traits, that is, color face, iris and eye along with fingerprints of right and left hands, handwriting, palm-print, gait (silhouettes) and wrist-vein for person authentication. The images of different biometric traits were taken from different standard databases such as FEI, UTIRIS, CASIA, IAM and CIE. This system is comprised of three different super-classifiers to individually perform person identification. The individual classifiers corresponding to each super-classifier in their turn identify different biometric features and their conclusions are integrated together in their respective super-classifiers. The decisions from individual super-classifiers are integrated together through a mega-super-classifier to perform the final conclusion using programming based boosting. The mega-super-classifier system using different super-classifiers in a compact form is more reliable than single classifier or even single super-classifier system. The system has been evaluated with accuracy, precision, recall and F-score metrics through holdout method and confusion matrix for each of the single classifiers, super-classifiers and finally the mega-super-classifier. The different performance evaluations are appreciable. Also the learning and the recognition time is fairly reasonable. Thereby making the system is efficient and effective.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.05a
• /
• pp.833-836
• /
• 2008

모바일 기기 보급이 급증함에 따라 모바일 환경에서 이루어지는 Peer-to-Peer(P2P)방식에 대한 연구가 활발히 진행되고 있다. 모바일 기기는 무선통신이라는 특성이 있기 때문에, 이동하는 peer의 에너지 절약이 항상 요구된다. Peer에 저장된 에너지를 모두 사용하게 되면 네트워크에 참여할 수가 없고, 에너지가 없는 peer 수가 지속적으로 증가되면 파일 탐색이 불가능하게 된다. Super peer를 기반으로 하는 계층적인 P2P 시스템은 기존의 단일 계층인 P2P 시스템보다 적은 양의 메시지를 이용하여 파일을 탐색할 수 있다는 장점이 있지만, super peer로 선정된 peer는 주변의 sub-peer들로부터 파일 탐색 요청 메시지를 지속적으로 받게 되어 빨리 에너지가 소모되는 단점이 있다. 본 논문은 super peer를 이용한 모바일 P2P 시스템에서 peer들의 남은 에너지를 고려하여 에너지가 적은 peer의 사용을 피하여 상대적으로 에너지가 많이 남은 peer들의 사용을 유도하는 시스템을 제안하였다. 실험을 통해서 성능평가를 하였으며, 제안된 시스템의 네트워크 유지시간이 기존 super peer를 이용한 시스템보다 오래 지속됨을 확인하였다.

• Journal of Korea Multimedia Society
• /
• v.11 no.12
• /
• pp.1706-1715
• /
• 2008

In a super-peer based P2P network, the network is clustered and each cluster is managed by a special peer, which is called a super-peer. A Super-peer has information of all the peers in its cluster. This type of clustered P2P model is known to have efficient information search and less traffic load than unclustered P2P model. In this paper, we compute the message traffic cost incurred by peers' query, join and update actions within a cluster as well as between the clusters. With these values, we estimate the optimal number of super-peers that minimizes the traffic cost for the various size of super-peer based P2P networks.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.382-387
• /
• 2023

현재 GPT-4와 같은 거대한 언어 모델이 기계 번역, 요약 및 대화와 같은 다양한 작업에서 압도적인 성능을 보이고 있다. 그러나 이러한 거대 언어 모델은 학습 및 적용에 상당한 계산 리소스와 도메인 특화 미세 조정이 어려운 등 몇 가지 문제를 가지고 있다. In-Context learning은 데이터셋에서 추출한 컨택스트의 정보만으로 효과적으로 작동할 수 있는 효율성을 제공하여 앞선 문제를 일부 해결했지만, 컨텍스트의 샷 개수와 순서에 민감한 문제가 존재한다. 이러한 도전 과제를 해결하기 위해, 우리는 Super In-Context Learning (SuperICL)을 활용한 새로운 방법론을 제안한다. 기존의 SuperICL은 적용한 플러그인 모델의 출력 정보를 이용하여 문맥을 새로 구성하고 이를 활용하여 거대 언어 모델이 더욱 잘 분류할 수 있도록 한다. Super In-Context Learning for Generation은 다양한 자연어 생성 작업에 효과적으로 최적화하는 방법을 제공한다. 실험을 통해 플러그인 모델을 교체하여 다양한 작업에 적응하는 가능성을 확인하고, 자연어 생성 작업에서 우수한 성능을 보여준다. BLEU 및 ROUGE 메트릭을 포함한 평가 결과에서도 성능 향상을 보여주며, 선호도 평가를 통해 모델의 효과성을 확인했다.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.65 no.1
• /
• pp.97-102
• /
• 2016

In this paper, a bidirectional power supply charger is proposed. This system used a solar cell panel to generate electricity and used super capacitors to store these energies, which can be used for the robot cleaner or some other electronic products. This system include a phase-shift controlled bidirectional dual active bridge (DAB) converter, solar panel super capacitors and DSP controller. In the daytime it can charge to the super capacitors to store the energy generated by the solar cell panel and in the night it will release the energy stored in the super capacitors to loads. A prototype of the proposed bidirectional power supply charger system was designed which can achieve 18V to 30V input, 10V/20W output to super capacitors and 9V/6.5W output when it acts as a charger for the robot cleaner. The system is verified to be sTable and reliable by both the simulation and experimental results.