• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06c
• /
• pp.197-200
• /
• 1998

본 논문에서는 유전자 알고리즘(genetic Algorithm)을 사용하여 벡터 양자화(vector quantization : VQ)를 수행하는 방법을 제안하고자 한다. 벡터 양자화를 수행하여 코드북(codebook)을 생성할 때 생성된 코드북과 학습벡터(training vector)사이에는 반드시 양자화 오차(quantization error)가 발생하는데 기존의 K-means 알고리듬을 사용하여 코드북을 생성했을 경우 양자화 오차를 줄이는데 한계가 있었다. 본 논문에서 제안하는 유전자 알고리즘을 이용한 벡터 양자화는 이 양자화 오차를 감소시키기 위해서 연구되었다. 제안한 방법의 성능을 평가하기 위해 음성데이터를 기존의 K-means 알고리즘에서 클러스터의 중심을 선택하는 방법중의 하나인 Minimax방법으로 코드북을 생성하여 제안한 방법과 양자화 오차를 비교한 결과 양자화 오차가 감소됨을 알 수 있었다.

• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.4
• /
• pp.499-505
• /
• 1998

In developing an automated surface inspect algorithm, we have designed a hierarchical classifier using neural network. The defects which exist on the surface of cold mill strip have a scattering or singular distribution. We have considered three major problems, that is preprocessing, feature extraction and defect classification. In preprocessing, Top-hit transform, adaptive thresholding, thinning and noise rejection are used Especially, Top-hit transform using local minimax operation diminishes the effect of bad lighting. In feature extraction, geometric, moment, co-occurrence matrix, and histogram ratio features are calculated. The histogram ratio feature is taken from the gray-level image. For defect classification, we suggest a hierarchical structure of which nodes are multilayer neural network classifiers. The proposed algorithm reduced error rate by comparing to one-stage structure.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.44 no.5
• /
• pp.94-102
• /
• 2007

In de-noising, it is wellknown that wavelet-thresholding algorithm shows near-optimal performances in the minimax sense. However, the wavelet-thresholding algorithm is difficult in realization it on hardware, such as FPGA, because of wavelet function complexity. In this paper, we present a new do-noising technique with the binary tree structured filter bank, which is based on the signal power ratio of each subbands to the total signal power. And we realize it on FPGA. For simple realization, the filter banks are designed by Hadamard transform coefficients. The simulation and hardware experimental results show that the performance of the proposed method is similar with that of soft thresholding de-noising algorithm based on wavelets, nevertheless it is simple.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.17 no.2
• /
• pp.131-136
• /
• 2011

The furrow calculation and traffic measurement of sailing ship using computer vision are useful methods to prevent maritime accident by predicting the possibility of an accident occurrence in advance. In this paper, sailing ships are recognized using image erosion, differential operator and minimax value, which can be verified directly because the calculated coordinates are displayed on electronic navigation chart. The developed algorithm based on area information of this paper has the advantage which is compared to the conventional radar system focused on point information.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1993.10a
• /
• pp.118-118
• /
• 1993

We consider a discrete event dynamic system called periodic job shop, where an identical mixture of items called minimal part set(MPS) is repetitively produced in the same processing order and the primary performance measure is the cycle time. The precedence relationships among events(starts of operations) are represented by a directed graph with rocurront otructure. When each operation starts as soon as all its preceding operations complete(called earliest starting), the occurrences of events are modeled in a linear system using a special algebra called minimax algebra. By investigating the eigenvalues and the eigenvectors, we develop conditions on the directed graph for which a stable steady state or a finite eigenvector exists. We demonstrate that each finite eigenvector, characterized as a finite linear combination of a class of eigenvalue, is the minimum among all the feasible schedules and an identical schedule pattern repeats every MPS. We develop an efficient algorithm to find a schedule among such schedules that minimizes a secondary performance measure related to work-in-process inventory. As a by-product of the linear system approach, we also propose a way of characterizing stable steady states of a class of discrete event dynamic systems.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.10c
• /
• pp.428-432
• /
• 2007

We present a novel method for determining k nearest neighbors, which accurately recognizes the underlying clusters in a data set. To this end, we introduce the "tiling neighborhood" which is constructed by tiling a number of small local circles rather than a single circle, as existing neighborhood schemes do. Then we formulate the problem of determining the tiling neighborhood as a minimax optimization, leading to an efficient message passing algorithm. For several real data sets, our method outperformed the k-nearest neighbor method. The results suggest that our method can be an alternative to existing for general classification tasks, especially for data sets which have many missing values.

• Journal of the Korean Data and Information Science Society
• /
• v.10 no.2
• /
• pp.289-298
• /
• 1999

We consider a minimax approach to make a design robust to many types or uncertainty arising in reality when dealing with non-normal linear models. We try to build a design to protect against the worst case, i.e. to improve the "efficiency" of the worst situation that can happen. In this paper, we especially deal with the generalized linear model. It is a known fact that the generalized linear model is a universal approach, an extension of the normal linear regression model to cover other distributions. Therefore, the optimal design for the generalized linear model has very similar properties as the normal linear model except that it has some special characteristics. Uncertainties regarding the unknown parameters, link function, and the model structure are discussed. We show that the suggested approach is proven to be highly efficient and useful in practice. In the meantime, a computer algorithm is discussed and a conclusion follows.

• Communications for Statistical Applications and Methods
• /
• v.27 no.4
• /
• pp.445-458
• /
• 2020

The least absolute shrinkage and selection operator (LASSO) is a popular method for a high-dimensional regression model. LASSO has high prediction accuracy; however, it also selects many irrelevant variables. In this paper, we consider the moderately clipped LASSO (MCL) for the high-dimensional generalized linear model which is a hybrid method of the LASSO and minimax concave penalty (MCP). The MCL preserves advantages of the LASSO and MCP since it shows high prediction accuracy and successfully selects relevant variables. We prove that the MCL achieves the oracle property under some regularity conditions, even when the number of parameters is larger than the sample size. An efficient algorithm is also provided. Various numerical studies confirm that the MCL can be a better alternative to other competitors.

• Journal of KIISE:Software and Applications
• /
• v.27 no.5
• /
• pp.543-550
• /
• 2000

An adaptive bit-reduced mean absolute difference (ABRMAD) is presented as a criterion for the block-matching algorithm (BMA) to reduce the complexity of the VLSI Implementation and to improve the processing time. The ABRMAD uses the lower pixel resolution of the significant bits instead of full resolution pixel values to estimate the motion vector (MV) by examining the pixels Ina block. Simulation results show that the 4-bit ABRMAD has competitive mean square error (MSE)results and a half less hardware complexity than the MAD criterion, It has also better characteristics in terms of both MSE performance and hardware complexity than the Minimax criterion and has better MSE performance than the difference pixel counting(DPC), binary block-matching with edge-map(BBME), and bit-plane matching(BPM) with the same number of bits.

• Journal of Satellite, Information and Communications
• /
• v.3 no.2
• /
• pp.48-53
• /
• 2008

In the future, satellite communication systems, such as ISDB in Japan and DVB in Europe, are required to support higher transmission date rate for providing multimedia services including HDTV, high rate data communication etc. Considering the effectiveness of OFDM technique in efficient usage of frequency bandwidth and its robustness to the multi-path fading, several OFDM based standards have been proposed for satellite communication. However, the problem of high Peak to Average Power Ratio is one of the main obstacles for the implementation of OFDM based system. Many PAPR reduction schemes have been proposed for OFDM systems. Among these, the partial transmit sequences (PTS) is attractive as they obtain better PAPR property by modifying OFDM signals without distortion. In this paper, considering the complexity issue, we present a simplified minimum maximum (minimax) criterion and Sub-Optimal PTS algorithm to optimize the phase factor. This algorithm can be dynamically made tradeoff`f between performance and complexity on demand. In addition, we integrate guided scrambling (GS) with this method. Simulation in multiple antenna based OFDM system proves that the proposed Hybrid schemes can get much more PAPR reduction and do not require transmission of side information (SI). Thus it is helpful when implementing OFDM technique in satellite communication system.