• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.2
• /
• pp.742-762
• /
• 2015

In this study, we propose a distance metric learning approach called discriminant metric learning (DML) for face verification, which addresses a binary-class problem for classifying whether or not two input images are of the same subject. The critical issue for solving this problem is determining the method to be used for measuring the distance between two images. Among various methods, the large margin nearest neighbor (LMNN) method is a state-of-the-art algorithm. However, to compensate the LMNN's entangled data distribution due to high levels of appearance variations in unconstrained environments, DML's goal is to penalize violations of the negative pair distance relationship, i.e., the images with different labels, while being integrated with LMNN to model the distance relation between positive pairs, i.e., the images with the same label. The likelihoods of the input images, estimated using DML and LMNN metrics, are then weighted and combined for further analysis. Additionally, rather than using the k-nearest neighbor (k-NN) classification mechanism, we propose a verification mechanism that measures the correlation of the class label distribution of neighbors to reduce the false negative rate of positive pairs. From the experimental results, we see that DML can modify the relation of negative pairs in the original LMNN space and compensate for LMNN's performance on faces with large variances, such as pose and expression.

• Journal of KIISE:Computer Systems and Theory
• /
• v.28 no.1_2
• /
• pp.64-72
• /
• 2001

이 논문은 행렬 탐색 방법을 이용하여 평면상의 η개의 점에 대한 모든 L$_{p}$, 1$\leq$P$\leq$$\infty$ 거리의 양방향 각도제한 근접 점 문제를 0(nlogn) 시간에 계산하는 알고리즘을 고안한다. 이 방법은 최적의 시간 복잡도를 가지며 궤적추적 법을 쓰지 않기 때문에 수치오차가 적으며 구현이 용이하고 실용적이다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10a
• /
• pp.542-544
• /
• 2000

이 논문은 행렬탐색 방법을 이용하여 평면상의 n 개의 점에 대한 Lp, 1$\leq$p$\leq$$\infty$ 거리의 양방향 각도제한 근접 점 문제를 O(nlogn) 시간에 계산하는 알고리즘을 고안한다. 이 방법은 최적의 시간 복잡도를 가지며 궤적추적 법을 쓰지 않기 때문에 구현이 용이하고 실용적이다.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2002.05a
• /
• pp.645-651
• /
• 2002

we propose a k-nearest neighbors(k-NN) based pattern selection method. The method tries to select the patterns that are near the decision boundary and that are correctly labeled. The simulations over synthetic data sets showed promising results: (1) By converting a non-separable problem to a separable one, the search for an optimal error tolerance parameter became unnecessary. (2) SVM training time decreased by two orders of magnitude without any loss of accuracy. (3) The redundant SVM were substantially reduced.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.04a
• /
• pp.1264-1266
• /
• 2009

We introduce a location-based routing model applied for grid backbone nodes in wireless mesh network. The number of paths with nearest distance between two nodes is calculated and used as key parameter to execute routing algorithm. Node will increase the transmission range that makes a trade off with data rate to reach its neighbors when node itself is isolated. The routing model is lightweight and oriented thanks to the simple but efficient routing algorithm.

• The Journal of the Korea Contents Association
• /
• v.9 no.11
• /
• pp.54-62
• /
• 2009

To avoid the redundant investment on the project selection process, it is necessary to check whether the submitted research topics have been proposed or carried out at other institutions before. This is possible through the search engines adopted by the keyword matching algorithm which is based on boolean techniques in national-sized research results database. Even though the accuracy and speed of information retrieval have been improved, they still have fundamental limits caused by keyword matching. This paper examines implemented TFIDF-based algorithm, and shows an experiment in search engine to retrieve and give the order of priority for similar and redundant documents compared with research proposals, In addition to generic TFIDF algorithm, feature weighting and K-Nearest Neighbors classification methods are implemented in this algorithm. The documents are extracted from NDSL(National Digital Science Library) web directory service to test the algorithm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.1C
• /
• pp.69-77
• /
• 2002

Sine digital signal processing for motion pictures requires huge amount of data computation to store, manipulate and transmit, more effective data compression is necessary. Therefore, the ITU-T recommended H.26x as data compression standards for digital motion pictures. The data compression method that eliminates time redundancies by motion estimation using relationship between picture frames has been widely used. Most video conding systems employ block matching algorithm for the motion estimation and compensation, and the algorithm is based on the minimun value of cast functions. Therefore, fast search algorithm rather than full search algorithm is more effective in real time low data rates encodings such as H.26x. In this paper, motion estimation employing the Nearest-Neighbors algorithm is designed to reduce search time using FPGA, coded in VHDL, and simulated and verified using Xilink Foundation.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.2
• /
• pp.20-28
• /
• 2003

Recently we need a faster and more accurate motion vector search algorithm for ASIC(Application Specific IC) or small systems. Block motion estimation using Full Search(FS) algorithm provides the best visual quality and PSNR, but it requires intensive computations. The previously proposed fast algorithms reduced the number of computations by limiting the number of searching locations. This is accomplished at the expense of less accuracy of motion estimation and gives rise to an appreciably higher SAD(Sum of Absolute Difference) for motion compensated images. In this paper we exploit the spatial correlation of motion vectors and present a fast motion estimation scheme which uses the predicted motion vector(PMV). The PMV scheme is more clear and simpler than the previously proposed algorithms which also use adjacent motion vectors. Simulation results with standard video sequences show that the PMV scheme is faster and more accurate than other algorithms such as Nearest-Neighbors Search(NNS) algorithm.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.40 no.4
• /
• pp.154-163
• /
• 2017

Collaborative filtering, one of the most widely used techniques to build recommender systems, is based on the idea that users with similar preferences can help one another find useful items. Credit card user behavior analytics show that most customers hold three or less credit cards without duplicates. This behavior is one of the most influential factors to data sparsity. The 'cold-start' problem caused by data sparsity prevents recommender system from providing recommendation properly in the personalized credit card recommendation scenario. We propose a personalized credit card recommender system to address the cold-start problem, using multiple user profiles. The proposed system consists of a training process and an application process using five user profiles. In the training process, the five user profiles are transformed to five user networks based on the cosine similarity, and an integrated user network is derived by weighted sum of each user network. The application process selects k-nearest neighbors (users) from the integrated user network derived in the training process, and recommends three of the most frequently used credit card by the k-nearest neighbors. In order to demonstrate the performance of the proposed system, we conducted experiments with real credit card user data and calculated the F1 Values. The F1 value of the proposed system was compared with that of the existing recommendation techniques. The results show that the proposed system provides better recommendation than the existing techniques. This paper not only contributes to solving the cold start problem that may occur in the personalized credit card recommendation scenario, but also is expected for financial companies to improve customer satisfactions and increase corporate profits by providing recommendation properly.

• Journal of the Korea Convergence Society
• /
• v.13 no.3
• /
• pp.33-41
• /
• 2022

In this article, we proposed to predict natural gas (NG) leakage levels through feature selection based on a factor analysis (FA) of the integrating the Korean Meteorological Agency data and natural gas leakage data for considering complex factors. The paper has been divided into three modules. First, we filled missing data based on the linear interpolation method on the integrated data set, and selected essential features using FA with OrdinalEncoder (OE)-based normalization. The dataset is labeled by K-means clustering. The final module uses four algorithms, K-nearest neighbors (KNN), decision tree (DT), random forest (RF), Naive Bayes (NB), to predict gas leakage levels. The proposed method is evaluated by the accuracy, area under the ROC curve (AUC), and mean standard error (MSE). The test results indicate that the OrdinalEncoder-Factor analysis (OE-F)-based classification method has improved successfully. Moreover, OE-F-based KNN (OE-F-KNN) showed the best performance by giving 95.20% accuracy, an AUC of 96.13%, and an MSE of 0.031.