• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.05a
• /
• pp.396-399
• /
• 2009

Stereo vision system is useful method for inferring 3D depth information from two or more images. So it has been the focus of attention in this field for a long time. Stereo matching is the process of finding correspondence points in two or more images. A central problem in a stereo matching is that it is difficult to satisfy both the computation time problem and accuracy at the same time. To resolve this problem, we proposed a new stereo matching technique using weighted cost aggregation. To begin with, we extract the weight in given stereo images based on features. We compute the costs of the pixels in a given window using correlation of weighted color, brightness and distance information. Then, we match pixels in a given window between the reference and target images of a stereo pair. To demonstrate the effectiveness of the algorithm, we provide experimental data from several synthetic and real scenes. The experimental results show the improved accuracy of the proposed method.

• Physical Therapy Korea
• /
• v.13 no.4
• /
• pp.10-15
• /
• 2006

This study examined the ability of human subjects to match a force in their quadriceps muscle during fatigue. Twenty subjects (mean age: 23.4 yrs, mean height: 167.8 cm, mean weight, 62.6 kg) were enrolled in the experiment. In the force-matching task, the quadriceps muscle generated 50% of the MVIC (maximum voluntary isometric contraction) torque under visual control and then without visual feedback. After inducing fatigue in the quadriceps muscle, the subjects were required to match 50% of the MVIC torque without visual feedback. The perceived magnitude of the force and force-matching errors were measured. 50% of the MVIC torque was perceived from 39.96 Nm in the pre-fatigue condition to 44.95 Nm in the post-fatigue condition. 50% of the MVIC torque-matching errors increased significantly from .55% in the pre-fatigue condition to 9.6% in the post-fatigue condition (p<.001). in addition, there were significantly more force-matching errors in women than in men (p<.01). In conclusion muscle fatigue can interfere with a subject's ability to match a force. This suggests that muscle fatigue may contributes to the sensitization of the proprioception.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.11
• /
• pp.68-78
• /
• 1999

A new algorithm for tracking feature points in an image sequence is presented. Most existing feature tracking algorithms often produce false trajectories, because the matching measures do not precisely reflect motion characteristics. In this paper, three attributes including spatial coordinate, motion direction and motion magnitude are used to calculate the feature point correspondence. The trajectories of feature points are determined by calculation the matching measure, which is defined as the minimum weighted Euclidean distance between two feature points. The weights of the attributes are updated reflecting the motion characteristics, so that the robust tracking of feature points is achieved. The proposed algorithm can find the trajectories correctly which has been shown by experimental results.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.841-842
• /
• 2008

스테레오 매칭(Stereo Matching) 기법에 대한 전역적인 방법과 지역적인 방법에 대한 연구가 활발하게 진행되고 있다. 최근의 적응적 영역 가중치 방법(Adaptive Support-Weight)은 매우 뛰어난 결과에 비해 많은 계산 시간이 필요하다. 따라서 로봇시스템에서 스테레오 매칭을 이용하기에는 부적합하다. 본 논문에서는 분리 가능한 Bilateral 필터를 이용하여 빠른 스테레오매칭 기법을 제안한다

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.604-606
• /
• 2004

클라이언트/서버 시스템은 클라이언트가 수행할 수 없거나 곤란한 프로그램을 서버에서 수행하여 클라이언트의 부하를 줄여준다. 하지만 서버의 리소스를 요청하는 클라이언트의 수가 많아지면 이미 많은 리소스를 사용하고 있는 서버에게 리소스를 요청하거나, 리소스를 사용하지 않는 서비스에게는 요청하지 않는 경우가 발생하여 서버에게 너무 많은 부하를 주거나 전혀 부하가 생기지 않을 수 있다. 이에 본 논문에서는 Jini 기반의 Agent Manager를 이용하여 서버와 클라이언트의 정보를 수집한 다음 MAUT와 Maximum-Weight Matching을 이용하여 클라이언트와 서버 간 연결의 전체적인 만족도를 높이는 Jini 기반의 최적의 컴퓨팅 리소스 제안 시스템을 연구하였다.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 1994.05a
• /
• pp.9-12
• /
• 1994

Epoxy-Ceramic Composite have good insulating, therma1 and mechanical properties, so it is studied actively on this material. In this thesis, we made a composite material b)\ulcorner filling Epoxy Resin with ceramics treated with Sillane Coupling Agent and studied dielectric and insulating characteristics according to treatment density of Sillane Coupling Agent and weight percent of filler. As a result, loss tangent increase and electrical breakdown voltage decrease according to increasing treatment density of sillane coupling agent because Interface matching between matrix and filler is not good. The best treatment density of sillane coupling agent is 0.5% water solution, in this density the best interface matching is achieved so good dielectric and insulation characteristics are shown. Dielectric and insulation characteristics according to weight percent of filler are best at 25wt.

• ETRI Journal
• /
• v.41 no.3
• /
• pp.371-382
• /
• 2019

Machine reading comprehension is the task of understanding a given context and finding the correct response in that context. A simple recurrent unit (SRU) is a model that solves the vanishing gradient problem in a recurrent neural network (RNN) using a neural gate, such as a gated recurrent unit (GRU) and long short-term memory (LSTM); moreover, it removes the previous hidden state from the input gate to improve the speed compared to GRU and LSTM. A self-matching network, used in R-Net, can have a similar effect to coreference resolution because the self-matching network can obtain context information of a similar meaning by calculating the attention weight for its own RNN sequence. In this paper, we construct a dataset for Korean machine reading comprehension and propose an $S^2-Net$ model that adds a self-matching layer to an encoder RNN using multilayer SRU. The experimental results show that the proposed $S^2-Net$ model has performance of single 68.82% EM and 81.25% F1, and ensemble 70.81% EM, 82.48% F1 in the Korean machine reading comprehension test dataset, and has single 71.30% EM and 80.37% F1 and ensemble 73.29% EM and 81.54% F1 performance in the SQuAD dev dataset.

• Journal of Biomedical Engineering Research
• /
• v.28 no.5
• /
• pp.601-608
• /
• 2007

2 Dimensional Gel Electrophoresis(2DGE) is an essentialmethodology for analysis on the expression of various proteins. For example, information for the location, mass, expression, size and shape of the proteins obtained by 2DGE can be used for diagnosis, prognosis and biological progress by comparison of patients with the normal persons. Protein spot matching for this purpose is comparative analysis of protein expression pattern for the 2DGE images generated under different conditions. However, visual analysis of protein spots which are more than several hundreds included in a 2DGE image requires long time and heavy effort. Furthermore, geometrical distortion makes the spot matching for the same protein harder. In this paper, an iterative algorithm is introduced for more efficient spot matching. Proposed method is first performing global matching step, which reduces the geometrical difference between the landmarks and the spot to be matched. Thus, movement for a spot is defined by a weighted sum of the movement of the landmark spots. Weight for the summation is defined by the inverse of the distance from the spots to the landmarks. This movement is iteratively performed until the total sum of the difference between the corresponding landmarks is larger than a pre-selected value. Due to local distortion generally occurred in 2DGE images, there are many regions in whichmany spot pairs are miss-matched. In the second stage, the same spot matching algorithm is applied to such local regions with the additional landmarks for those regions. In other words, the same method is applied with the expanded landmark set to which additional landmarks are added. Our proposed algorithm for spot matching empirically proved reliable analysis of protein separation image by producing higher accuracy.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.10
• /
• pp.123-131
• /
• 1999

In this paper, we proposed stereo matching algorithm by dynamic programming with edges emphasized. Existing algorithms show blur generally at depth discontinuities owing to smoothness constraint and non-existence of matching pixel in occlusion regions. Also it accompanies matching error by lackness of matching information in the untextured regions. This paper defines new cost function to make up for the problems occurred to existing algorithms. It is possible through deriving matching of edges in left and right images to be carried out between edge regions anf deriving that in the other regions to be peformed between the other regions. In case of the possibility that edges can be Produced in a large amount, matching between edge information adds weight to cost function in proportion to Path distance. Proposed algorithm was applied to various images obtained by convergent camera model as well as parallel camera model. As the result, proposed algorithm showed improved performance in the aspect of matching error and processing in the occlusion regions compared to existing algorithms. Also it could improve blur especially in discontinuity regions.

• The Journal of the Korea Contents Association
• /
• v.15 no.4
• /
• pp.1-8
• /
• 2015

This paper presents a Electrocardiogram(ECG) identification method using adaptive weight based on Least Mean Square Error(LMSE) optimization. With a preprocessing for noise suppression, we extracts the average ECG signal and its standard deviation at every time instant. Then the extracted information is stored in database. ECG identification is achieved by matching an input ECG signal with the information in database. In computing the matching scores, the standard deviation is used. The scores are computed by applying adaptive weights to the values of the input signal over all time instants. The adaptive weight consists of two terms. The first term is the inverse of the standard deviation of an input signal. The second term is the proportional one to the standard deviation between user SAECGs stored in the DB. Experimental results show up to 100% recognition rate for 32 registered people.