• Journal of Sensor Science and Technology
• /
• v.21 no.1
• /
• pp.59-67
• /
• 2012

It was reported that the latest sensor technology allow an 65536 conductive polymer sensor array to be made with broad but overlapping selectivity to different families of chemicals emulating the characteristics found in biological olfaction. However, the supernumerary redundancy always accompanies great error and risk as well as an inordinate amount of computation time and local minima in signal processing, e.g. neural networks. In this paper, we propose a new method to reduce the number of sensor for analysis by reducing redundancy between sensors and by removing unstable sensors using the cosine similarity method and to decide on representative sensor using FCM(Fuzzy C-Means) algorithm. The representative sensors can be just used in analyzing. And, we introduce DWT(Discrete Wavelet Transform) for data compression in the time domain as preprocessing. Throughout experimental trials, we have done a comparative analysis between gas sensor data with and without reduced redundancy. The possibility and superiority of the proposed methods are confirmed through experiments.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.37 no.6
• /
• pp.551-558
• /
• 2013

A support vector machine (SVM) is a very powerful classification algorithm when a set of training data, each marked as belonging to one of several categories, is given. Therefore, SVM techniques have been used as one of the diagnostic tools in machine learning as well as in pattern recognition. In this paper, we present the results of classifying ball bearing fault types and severities using SVM with an optimized feature set based on the minimum distance rule. A feature set as an input for SVM includes twelve time-domain and nine frequency-domain features that are extracted from the measured vibration signals and their decomposed details and approximations with discrete wavelet transform. The vibration signals were obtained from a test rig to simulate various bearing fault conditions.

• Wind and Structures
• /
• v.31 no.3
• /
• pp.269-285
• /
• 2020

The statistical characteristics of typhoon wind speed records tend to have a considerable time-varying trend; thus, the stationary wind model may not be appropriate to estimate the wind characteristics of typhoon events. Several nonstationary wind speed models have been proposed by pioneers to characterize wind characteristics more accurately, but comparative studies on the applicability of the different wind models are still lacking. In this study, three landfall typhoons, Ampil, Jongdari, and Rumbia, recorded by ultrasonic anemometers atop the Shanghai World Financial Center (SWFC), are used for the comparative analysis of stationary and nonstationary wind characteristics. The time-varying mean is extracted with the discrete wavelet transform (DWT) method, and the time-varying standard deviation is calculated by the autoregressive moving average generalized autoregressive conditional heteroscedasticity (ARMA-GARCH) model. After extracting the time-varying trend, the longitudinal wind characteristics, e.g., the probability distribution, power spectral density (PSD), turbulence integral scale, turbulence intensity, gust factor, and peak factor, are comparatively analyzed based on the stationary wind speed model, time-varying mean wind speed model and time-varying standard deviation wind speed model. The comparative analysis of the different wind models emphasizes the significance of the nonstationary considerations in typhoon events. The time-varying standard deviation model can better identify the similarities among the different typhoons and appropriately describe the nonstationary wind characteristics of the typhoons.

• Journal of Sensor Science and Technology
• /
• v.22 no.2
• /
• pp.162-173
• /
• 2013

The purpose of this paper is to classify VOC gases by emulating the characteristics found in biological olfaction. For this purpose, we propose new signal processing method based a polymeric chemical sensor array consisting of 4096 sensors which is created by NEUROCHEM project. To remove unstable sensors generated in the manufacturing process of very large scaled chemical sensor array, we used discrete wavelet transformation and cosine similarity. And, to remove the supernumerary redundancy, we proposed the method of selecting candidates of representative sensor representing sensors with similar features by Fuzzy c-means algorithm. In addition, we proposed an improved algorithm for selecting representative sensors among candidates of representative sensors to better enhance classification ability. However, Classification for very large scaled sensor array has a great deal of time in process of learning because many sensors are used for learning though a redundancy is removed. Throughout experimental trials for classification, we confirmed the proposed method have an outstanding classification ability, at transient state as well as steady state.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.9
• /
• pp.4389-4411
• /
• 2018

High dynamic range (HDR) images are becoming pervasive due to capturing or rendering of a wider range of luminance, but their special display equipment is difficult to be popularized because of high cost and technological problem. Thus, HDR images must be adapted to the conventional display devices by applying tone mapping (TM) operation, which puts forward higher requirements for intellectual property protection of HDR images. As the robustness presents regional diversity in the low dynamic range (LDR) watermarked image after TM, which is different from the traditional watermarking technologies, a concept of watermarking activity is defined and used to distinguish the essential distinction of watermarking between LDR image and HDR image in this paper. Then, a novel robust HDR image watermarking algorithm is proposed against TM operations. Firstly, based on the hybrid processing of redundant discrete wavelet transform and singular value decomposition, the watermark is embedded by modifying the structure information of the HDR image. Distinguished from LDR image watermarking, the high embedding strength can cause more obvious distortion in the high brightness regions of HDR image than the low brightness regions. Thus, a perceptual brightness mask with low complexity is designed to improve the imperceptibility further. Experimental results show that the proposed algorithm is robust to the existing TM operations, with taking into account the imperceptibility and embedded capacity, which is superior to the current state-of-art HDR image watermarking algorithms.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.12C
• /
• pp.1193-1200
• /
• 2005

In this paper, we proposed an image hiding method which decreases the amount of calculation encrypting partial data rather than the whole image data using a discrete wavelet transform and a linear scalar quantization which have been adopted as the main technique in JPEG2000 standard and then implemented the proposed algorithm to hardware. A chaotic system was used instead of encryption algorithms to reduce further amount of calculation. It uses a method of random changing method using the chaotic system of the data in a selected subband. For ciphering the quantization index it uses a novel image encryption algorithm of cyclical shifting to the right or left direction and encrypts two quantization assignment method (Top-down coding and Reflection coding), made change of data less. The experiments have been performed with the proposed methods implemented in software for about 500 images. The hardware encryption system was synthesized to find the gate-level circuit with the Samsung $0.35{\mu}m$ Phantom-cell library and timing simulation was performed, which resulted in the stable operation in the frequency above 100MHz.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.6
• /
• pp.27-32
• /
• 2005

An effective video watermarking algorithm is proposed to protect the copyright. The watermarking procedure is based on a three-dimensional discrete wavelet transform (3D DWT) and spread spectrum sequences. Two perceptual binary watermarks are preprocessed using mixing and pseudorandom permutation. After dividing the video sequence into video shots, the 3D DWT is performed, then the preprocessed watermarks are embedded into the 3D DWT coefficients, while considering robustness and invisibility, using two spread spectrum sequences defined as the user key. Experimental results show that the watermarked frames are subjectively indistinguishable from the original frames, plus the proposed video watermarking algorithm is sufficiently robust against such attacks as low pass filtering, frame dropping, frame average, and MPEG coding.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38A no.2
• /
• pp.164-173
• /
• 2013

In this paper, we propose multi-frame based super resolution algorithm by using motion vector normalization and edge pattern analysis. Existing algorithms have constraints of sub-pixel motion and global translation between frames. Thus, applying of algorithms is limited. And single-frame based super resolution algorithm by using discrete wavelet transform which robust to these problems is proposed but it has another problem that quantity of information for interpolation is limited. To solve these problems, we propose motion vector normalization and edge pattern analysis for 2*2 block motion estimation. The experimental results show that the proposed algorithm has better performance than other conventional algorithms.

• Journal of KIISE:Software and Applications
• /
• v.33 no.11
• /
• pp.936-941
• /
• 2006

To apply object recognition techniques to real environment, illumination compensation method should be developed. As effective illumination compensation model, we focused our attention on Retinex model and illumination-Reflectance model, implemented them, and experimented on their performance. We implemented Retinex model with Single Scale Retinex, Multi-Scale Retinex, and Retinex Neural Network and Multi-Scale Retinex Neural Network, neural network model of Retinex model. Also, we implemented illumination-Reflectance model with reflectance image calculation by calculating an illumination image by low frequency filtering in frequency domain of Discrete Cosine Transform and Wavelet Transform, and Gaussian blurring. We compare their illumination compensation performance to facial images under nine illumination directions. We also compare their performance after post processing using Principal Component Analysis(PCA). As a result, illumination Reflectance model showed better performance and their overall performance was improved when illumination compensated images were post processed by PCA.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.8B
• /
• pp.1586-1592
• /
• 1999

In DWMT, CMFB is employed in the synthesis/analysis part. The CMFB uses filters of greater length than the DFT, resulting in reduced interference between the carriers. In addition, the CMFB system is computationally efficient and fast algorithms are available for their implementation. Traditional designs for the prototype filters of CMFB usually involve nonlinear optimizations. Thus the required design time is considerably large even for small filter orders. In this paper, a prototype filter design method for CMFB is presented using optimal window method. The design process is reduced to the optimization of a single parameter and consequently the required design time is much less than those of the existing methods. It is shown that the stopband performance of the proposed method is better than that of the Kaiser window method.