• ETRI Journal
• /
• v.38 no.2
• /
• pp.366-375
• /
• 2016

This paper describes an algorithm to suppress composite noise in a two-microphone speech enhancement system for robust hands-free speech communication. The proposed algorithm has four stages. The first stage estimates the power spectral density of the residual stationary noise, which is based on the detection of nonstationary signal-dominant time-frequency bins (TFBs) at the generalized sidelobe canceller output. Second, speech-dominant TFBs are identified among the previously detected nonstationary signal-dominant TFBs, and power spectral densities of speech and residual nonstationary noise are estimated. In the final stage, the bin-wise output signal-to-noise ratio is obtained with these power estimates and a Wiener post-filter is constructed to attenuate the residual noise. Compared to the conventional beamforming and post-filter algorithms, the proposed speech enhancement algorithm shows significant performance improvement in terms of perceptual evaluation of speech quality.

• Journal of Magnetics
• /
• v.6 no.1
• /
• pp.27-30
• /
• 2001

We present a new procedure to evaluate the residual life of 1Cr-1Mo-0.25V steel from reversible magnetic permeability. The method is based on the existence of first harmonics in the differential magnetization around the coercive force. The apparatus is based on the detection of the voltage induced in a coil using a lock-in amplifier tuned to the frequency of the exciting coil. Results obtained for the first harmonics and Vickers hardness on aged samples show that the peak interval of reversible permeability and Vickers hardness decrease as aging time increases. The correlation between Vickers hardness and the peak interval of the reversible permeability may be used to evaluate the residual life of 1Cr-lMo-0.25V steel, nondestructively.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2000.04a
• /
• pp.270-276
• /
• 2000

Hard turning has the potential to replace grinding process and to achieve significant reduction in production time and cost. The main applications for hard turning is finishing process, namely grinding process. Therefore, it must be able to satisfy high surface integrity of the workpiece. This paper discusses surface quality in terms of residual stress and damaged layer with respect to cutting parameters in an intermittent hard turning. Damaged layer experiment is carried out orthogonal array. From that is based on the orthogonal array. From the response table, cutting parameters are analyzed from the view point of the damaged layer and residual stress. From this experimental results, even though in the intermittent hard turning, surface integrity turns out be good enough for replacing grinding process.

• Journal of Welding and Joining
• /
• v.19 no.2
• /
• pp.215-220
• /
• 2001

In the T-joint welding, the complete penetration joint which is obtained by groove welding with edge preparation is generally required thor the safety and reliability of structures but this way have the some defects such as increase of working time, consumed welding electrode quantity and large welding deformation. If there is no probrem, in the strength, T-joint welding without edge preparation will be profitably understood in the economical and welding deformation side. In this paper, we performed the finite element analysis to understand the characteristics of welding residual stresses on two models, complete penetration joint have the edge preparation and incomplete penetration joint without edge preparation, respectively. Especially, we observed the relation between welding residual stress distributed on the notch of gap in the root and external force in the incomplete penetration joint without edge preparation.

• Journal of the Korean Society for Precision Engineering
• /
• v.19 no.1
• /
• pp.135-142
• /
• 2002

Input shaping is a method for reducing residual vibration in computer controlled machines. Vibration is eliminated by convolving a sequence of impulses, an input shaper, with a desired system command to produce a shaped input. This paper shows the shape of sensitivity curve of input shaper as impulse interval T and analysis of robustness for input shaper on the z-plane. And a method is presented for designing equivalent input shaper considering sampling time $T_s$. And then we applied equivalent input shaper to crane system.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2005.05a
• /
• pp.51-54
• /
• 2005

Reinforced Concrete structures have been commonly regarded as fire-resisting constructions. In the case of high-strength concrete, however, the behavior of a concrete member under fire and after fire has characteristics in different way with normal strength concrete members because of spalling. The resonable evaluation about the residual strength and stiffness of members as well as material properties has to be conducted before reusing the fire-damaged structures or retrofitting or strengthening them. Therefore, the guideline is needed for evaluation the residual strength and stiffness. In this study, the fire test is conducted with parameters like concrete strength, fire time and cover thickness, etc. The loads-deflection curves are used for comparison and analysis with the parameters.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.483-484
• /
• 2007

In this paper, we describe a chip design technique for instrumentation amplifier using a nested-chopping technique. Conventional chopping technique uses a pair of chopper, but nested chopping technique uses two pairs of chopper to reduce residual offset and 1/f noise. The inner chopper pair removes the 1/f noise, while the outer chopper pair reduces the residual offset. Our instrumentation amplifier using a nested chopping technique has residual offset under 100 nV. We also implement very low frequency filter. Since this filter needs very large RC time constant, we use a technique named 'diode connected PMOS' to increase R with small die area. The total power consumption is 3.1 mW at the supply voltage of 3.3V with the 0.35um general CMOS technology. The die area of implemented chip was $530um{\times}300um$.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.163-165
• /
• 2019

The task of image denoising is gaining popularity in the computer vision research field. Its main objective of restoring the sharp image from given noisy input is demanded in all image processing procedure. In this work, we treat the process of residual hair removal on faces images similar to the task of image denoising. In particular, our method removes the residual hair that presents on the frontal or profile face images and in-paints it with the relevant skin color. To achieve this objective, we employ a deep neural network that able to perform both tasks in one time. Furthermore, simple technic of residual hair color augmentation is introduced to increase the number of training data. This approach is beneficial for improving the robustness of the network. Finally, we show that the experimental results demonstrate the superiority of our network in both quantitative and qualitative performances.

• Journal of Information Processing Systems
• /
• v.20 no.1
• /
• pp.13-23
• /
• 2024

Music brings pleasure and relaxation to people. Therefore, it is necessary to classify musical genres based on scenes. Identifying favorite musical genres from massive music data is a time-consuming and laborious task. Recent studies have suggested that machine learning algorithms are effective in distinguishing between various musical genres. However, meeting the actual requirements in terms of accuracy or timeliness is challenging. In this study, a hybrid machine learning model that combines a deep residual auto-encoder (DRAE) and support vector machine (SVM) for musical genre recognition was proposed. Eight manually extracted features from the Mel-frequency cepstral coefficients (MFCC) were employed in the preprocessing stage as the hybrid music data source. During the training stage, DRAE was employed to extract feature maps, which were then used as input for the SVM classifier. The experimental results indicated that this method achieved a 91.54% F1-score and 91.58% top-1 accuracy, outperforming existing approaches. This novel approach leverages deep architecture and conventional machine learning algorithms and provides a new horizon for musical genre classification tasks.

• Journal of Chest Surgery
• /
• v.35 no.8
• /
• pp.580-583
• /
• 2002

Background: Obtaining precise hemodynamic and morphological information in the early postoperative period after surgical correction of congenital heart disease is important in determining the need for future medical or surgical intervention. We investigated the residual shunting after surgical repair of simple ventricular septal defect in order to know the incidence of residual shunting in the postoperative period and the natural history of small residual shunts located in the peripatch area. Material and Method: Forty three consecutive patients under one year of age who underwent patch repair of a simple ventricular septal defect were evaluated for incidence of residual shunts by echocardiography. Result: Eleven patients had echocardiographic residual shunt in the peripatch area at immediate postoperative period, however, there were no patients who needed reoperation due to deteriorated hemodynamic effect of residual shunt. The incidence of residual shunts was not significantly different with type of ventricular septal defect and material used for closure. During follow up period, two patients were lost and remaining nine patients no longer showed evidence of residual shunt. The mean time of last evidence of shunt was $4.2{\pm}3.6$ months after operation. Conclusion: Residual peripatch shunt flow was frequently noted in the immediate postoperative period following surgical repair of ventricular septal defect, however, most of them were disappeared within six months.