• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.12A
• /
• pp.1238-1243
• /
• 2007

The computational complexity for the zero-forcing (ZF)-based multiuser (MU) multiple-input multiple-output(MIMO) preprocessing matrices can be immoderately large as the number of transmit antennas or users increases. In this paper, we show that the span of singular vector space of a matrix can be obtained from the singular vectors of the parted rows of that matrix with computational saving and propose a computationally efficient recursive-algorithm for achieving the ZF-based preprocessing matrices. Analysis about the complexities shows that a new recursive-algorithm can lighten the computational load.

• Journal of the Korea Convergence Society
• /
• v.11 no.10
• /
• pp.45-51
• /
• 2020

Recently, many games provide data related to the users' game play, and there have been a few studies that predict opponent move by combining machine learning methods. This study predicts opponent move using match data of a real-time strategy game named ClashRoyale and a multi-label classification based on machine learning. In the initial experiment, binary card properties, binary card coordinates, and normalized time information are input, and card type and card coordinates are predicted using random forest and multi-layer perceptron. Subsequently, experiments were conducted sequentially using the next three data preprocessing methods. First, some property information of the input data were transformed. Next, input data were converted to nested form considering the consecutive card input system. Finally, input data were predicted by dividing into the early and the latter according to the normalized time information. As a result, the best preprocessing step was shown about 2.6% improvement in card type and about 1.8% improvement in card coordinates when nested data divided into the early.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.4
• /
• pp.569-577
• /
• 2018

In this paper, with the aid of information which is included within data, preprocessing algorithm-based black plastic classifier is designed. The slope and area of spectrum obtained by using laser induced breakdown spectroscopy(LIBS) are analyzed for each material and its ensuing information is applied as the input data of the proposed classifier. The slope is represented by the rate of change of wavelength and intensity. Also, the area is calculated by the wavelength of the spectrum peak where the material property of chemical elements such as carbon and hydrogen appears. Using informations such as slope and area, input data of the proposed classifier is constructed. In the preprocessing part of the classifier, Principal Component Analysis(PCA) and fuzzy transform are used for dimensional reduction from high dimensional input variables to low dimensional input variables. Characteristic analysis of the materials as well as the processing speed of the classifier is improved. In the condition part, FCM clustering is applied and linear function is used as connection weight in the conclusion part. By means of Particle Swarm Optimization(PSO), parameters such as the number of clusters, fuzzification coefficient and the number of input variables are optimized. To demonstrate the superiority of classification performance, classification rate is compared by using WEKA 3.8 data mining software which contains various classifiers such as Naivebayes, SVM and Multilayer perceptron.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.718-723
• /
• 1993

In this paper, we propose a neural network with fuzzy preprocessor not only for improving the classification accuracy but also for being able to classify objects whose attribute values do not have clear boundaries. The fuzzy input signal representation scheme is included as a preprocessing module. It transforms imprecise input in linguistic form and precisely stated numerical input into multidimensional numerical values. The transformed input is processed in the postprocessing module. The experimental results indicate the superiority of the backpropagation network with fuzzy preprocessor in comparison to the conventional backpropagation network.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.2C
• /
• pp.147-161
• /
• 2003

Video object segmentation is one of the core technologies for content-based real-time MPEG-4 encoding system. For real-time requirement, the segmentation algorithm should be fast and accurate but almost all existing algorithms are computationally intensive and not suitable for real-time applications. The MPEG-4 VM(Verification Model) has provided basic algorithms for MPEG-4 encoding but it has many limitations in practical software development, real-time camera input system and compression efficiency. In this paper, we implemented the preprocessing system for real-time camera input and VOP extraction for content-based video coding and also implemented motion detection to achieve the 180 : 1 compression rate for real-time and high compression MPEG-4 encoding.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.1799-1802
• /
• 2003

This paper has been studied a Adaptive feature-factors based fingerprints recognition in many biometrics. we study preprocessing and matching method of fingerprints image in various circumstances by using optical fingerprint input device. The Fingerprint Recognition Technology had many development until now. But, There is yet many point which the accuracy improves with operation speed in the side. First of all we study fingerprint classification to reduce existing preprocessing step and then extract a Feature-factors with direction information in fingerprint image. Also in the paper, we consider minimization of noise for effective fingerprint recognition system.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.49.4-49
• /
• 2001

Digit recognition based on backpropagation neural networks, as an important application of pattern recognition, was attracted much attention. Although it has the advantages of parallel calculation, high error-tolerance, and learning capability, better recognition effects can only be achieved with some specific fixed format input of the digit image. Therefore, digit image preprocessing ability directly affects the accuracy of recognition. Here using Matlab software, the digit image was enhanced by resizing and neutral-rotating the extracted digit image, which improved the digit recognition capability of the backpropagation neural network under practical conditions. This method may also be helpful for recognition of other patterns with backpropagation neural networks.

• 제어로봇시스템학회:학술대회논문집
• /
• 1998.10a
• /
• pp.201-204
• /
• 1998

This paper presents a new construction method of candidate controllers using Multi-modal Neural Network(MNN). To improve a control performance of multiple controller, we construct, candidate controllers which consist of MNN. MNN can learn more complicated function than multilayer neural network. MNN consists of preprocessing module and neural network module. The preprocessing module transforms input signals into spectra which are used as input of the following neural network module. We apply the proposed method to multiple control system which controls the cart-pole balancing system and show the effectiveness of the proposed method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.3
• /
• pp.227-230
• /
• 2002

An important preprocessing for fingerprint recognition is the binarization operation, which takes as an input gray-scale image and returns a binary image as the output. The difficult in performing binarization is to find an appropriate threshold value. This paper presents a new adaptive binarization method, which determines the threshold value according to the brightness of local ridges and valleys. We experimentally show that the presented method results in better performance than a traditional method.

• Smart Media Journal
• /
• v.10 no.3
• /
• pp.9-22
• /
• 2021

Basically, machine learning models use input data to produce results. Sometimes, the input data is too complicated for the models to learn useful patterns. Therefore, feature engineering is a crucial data preprocessing step for constructing a proper feature set to improve the performance of such models. One of the most efficient methods for automating feature engineering is the autoencoder, which transforms the data from its original space into a latent space. However certain factors, including the datasets, the machine learning models, and the number of dimensions of the latent space (denoted by k), should be carefully considered when using the autoencoder. In this study, we design a framework to compare two data preprocessing approaches: with and without autoencoder and to observe the impact of these factors on autoencoder. We then conduct experiments using autoencoders with classifiers on popular datasets. The empirical results provide a perspective regarding the best suited autoencoder for these factors.