• Speech Sciences
• /
• v.10 no.2
• /
• pp.77-84
• /
• 2003

In this research, we seek the beginning of the speech and detect the stationary speech region using lip information. Performing running average of the estimated speech signal in the stationary region, we reduce the effect of musical noise which is inherent to the conventional MlMSE (Minimum Mean Square Error) speech enhancement algorithm. In addition to it, SFM (Spectral Flatness Measure) is incorporated to reduce the speech signal estimation error due to speaking habit and some lacking lip information. The proposed algorithm with Wiener filtering shows the superior performance to the conventional methods according to MOS (Mean Opinion Score) test.

• CDE review
• /
• v.22 no.2
• /
• pp.6-10
• /
• 2016

This paper presents the kinematics of a walking robot leg based on Jansen mechanism. By using simple mathematics, all trajectories of walking robot leg links can be calculated. A foot point trajectory is used to evaluate the performance of a walking robot leg. Trial and Error method is used to find a best combination of link lengths under certain restrictions. All simulations are performed by Matlab. Ground score, drag score, step size, foot lift, instant speed, and average speed of foot point trajectories are used for selecting the best one.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 1990.10a
• /
• pp.79-84
• /
• 1990

The modal parameter estimations of linear multi-degree-of-freedom structural dynamic systems are carried out in time domain. For this purpose, the equation of motion is transformed into the autoregressive and moving average model with auxiliary stochastic input (ARMAX) model. The parameters of the ARMAX model are estimated by using the sequential prediction error method. Then, the modal parameters of the system are obtained thereafter. Experimental results are given for a 3-story building model subject to ground exitations.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.7
• /
• pp.49-55
• /
• 2016

In digital forensic images, there is a serious problem that is distributed with various image types. For the problem solution, this paper proposes a classification algorithm of the forensic image types. The proposed algorithm extracts the 21-dim. feature vector with the contrast and energy from GLCM (Gray Level Co-occurrence Matrix), and the entropy of each image type. The classification test of the forensic images is performed with an exhaustive combination of the image types. Through the experiments, TP (True Positive) and FN (False Negative) is detected respectively. While it is confirmed that performed class evaluation of the proposed algorithm is rated as 'Excellent(A)' because of the AUROC (Area Under Receiver Operating Characteristic Curve) is 0.9980 by the sensitivity and the 1-specificity. Also, the minimum average decision error is 0.1349. Also, at the minimum average decision error is 0.0179, the whole forensic image types which are involved then, our classification effectiveness is high.

• Journal of Korea Multimedia Society
• /
• v.13 no.11
• /
• pp.1698-1705
• /
• 2010

In this paper, we propose a fast liver segmentation method from magnetic resonance(MR) images. Our method efficiently divides a MR image into a set of discrete objects, and boundaries based on the normalized gradient magnitude information. Then, the objects belonging to the liver are detected by using 2D seeded region growing with seed points, which are extracted from the segmented liver region of the slice immediately above or below the current slice. Finally, rolling ball algorithm, and connected component analysis minimizes false positive error near the liver boundaries. Our method was validated by twenty data sets and the results were compared with the manually segmented result. The average volumetric overlap error was 5.2%, and average absolute volumetric measurement error was 1.9%. The average processing time for segmenting one data set was about three seconds. Our method could be used for computer-aided liver diagnosis, which requires a fast and accurate segmentation of liver.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.4
• /
• pp.1248-1253
• /
• 2010

In this paper, the compensation algorithm for localization using the concept of equivalent distance rate(AEDR) in order to compensate ranging error in the SDS-TWR(Symmetric Double-Sided Two-Way Ranging) is proposed and the performance of the proposed algorithm is analyzed by the localization experiments. The ranging error of the SDS-TWR in the distance between mobile node and beacon node is measured to average 1m~8m by ranging experiments. But it is confirmed that the performance of the localization by the AEDR is better than that of the SDS-TWR 4 times in university auditorium and corridor, and the localization error of above 3~10m is reduced to average 2m and that of below 3m is reduced to average 1m respectively. It is concluded that the AEDR is superior to the NLOS(Non Line Of Sight) than LOS(Line Of Sight) in performance of ranging compensation for localization, and the AEDR is more helpful to localization systems practically considering the environment of sensor networks is under NLOS.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.10
• /
• pp.5215-5220
• /
• 2013

Recent construction in the field of 3D aversion is increasing interest in automation. This study is results about survey of the coastal landfill using automatic VRS-GPS surveying system. GPS is made with GRXI and SHC250 controller. Automatic surveying system is composed of DPS module, geomagnetism sensor, bluetooth, gimbals, IMU, etc and enables an automatic driving via entered into a route of position. The developed auto surveying system has installed the front and camera for vertical axis and can grasp situation of surveying with smartphone in real time. The comparative result between surveyed result with repetition method auto VRS-GPS surveying system observed surveyed result with VRS-RTK has shown that average error of x-axis is 0.009m, average error of y-axis, 0.010m and average error of height, 0.002m. This possibility was confirmed that field application.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.10A
• /
• pp.1034-1041
• /
• 2007

In this paper, we propose a FD USE (frequency domain minimum mean square error) equalizer algorithm for MBOK DS-UWB (M-ary bi-orthogonal keying direct sequence UWB) systems considered as a PHY proposal for high-speed wireless communication in IEEE 802.15.TG3a. The conventional FD MMSE equalization scheme has a structural limit due to insertion of the cyclic prefix (CP) in all transmit packets, but the proposed scheme is able to equalize the channel effect without CP. In order to overcome channel estimation error by multipath delay, we introduce a moving FFT and a moving average scheme. Compared with conventional FD MMSE equalizer and the traditional TD (time domain) MMSE-RAKE receiver, the proposed FD MMSE equalizer has better BER performance and we demonstrate this result by computer simulation.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.30 no.4
• /
• pp.302-310
• /
• 2010

In this paper, we studied Bagging neural network for predicting defect size of steam generator(SG) tube in nuclear power plant. Bagging is a method for creating an ensemble of estimator based on bootstrap sampling. For predicting defect size of SG tube, we first generated eddy current testing signals for 4 defect patterns of SG tube with various widths and depths. Then, we constructed single neural network(SNN) and Bagging neural network(BNN) to estimate width and depth of each defect. The estimation performance of SNN and BNN were measured by means of peak error. According to our experiment result, average peak error of SNN and BNN for estimating defect depth were 0.117 and 0.089mm, respectively. Also, in the case of estimating defect width, average peak error of SNN and BNN were 0.494 and 0.306mm, respectively. This shows that the estimation performance of BNN is superior to that of SNN.

• KIISE Transactions on Computing Practices
• /
• v.21 no.1
• /
• pp.19-28
• /
• 2015

In this paper, we proposed an algorithm for the identification of relieving or worsening current traffic congestion using historic traffic congestion patterns. Historical congestion patterns were placed in an adjacency list. The patterns were constructed to represent spatial and temporal length for status of a congested road. Then, we found information about historical traffic congestions that were similar to today's traffic congestion and will use that information to show how to change traffic congestion in the future. The most similar pattern to current traffic status among the historical patterns corresponded to starting section of current traffic congestion. One of our experiment results had average error when we compared identified changes of the congestion for one of the sections in the congestion road by using our proposal and real traffic status. The average error was 15 minutes. Another result was for the long congestion road consisting of several sections. The average error for this result was within 10 minutes.