• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.13 no.6
• /
• pp.490-500
• /
• 1988

In recent years the acquisition, processing and interpretation of three dimensional seisimic data, for the purpose of locating gas and reservoirs, have become practical. This paper exlores one way in which the volume data can be searched and visualized, which may aid the interpreter. The illusion of looking at a three dimensional volume can be obrained by fusing a stereoscopic pair of pictures. Each picture can be made by projecting each data point of the volume into a plane from a point where the eye is placed. The data valuse along any projection line can be summed to form the picture, or only a segment along the line can be selected. By selective projection, the volume can be searched and obscuring layers removed. The stereoscopic pictures show the physical models in there ture spatial positions. Projection of the envelope function of the seismic traces is shown to give improved depth perception compared with projection of the position amplitudes.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.11
• /
• pp.63-66
• /
• 1997

With the advent of high speed digital signal processing chips, new digital techniques have been introduced to the hearing instrument. This advanced hearing instrument circuitry has led to the need or and the development of new fitting approach. A number of different fitting approaches have been developed over the past few years, yet there has been little agreement on which approach is the "best" or most appropriate to use. However, when we develop not only new hearing aid, but also its fitting method, the intensive subject-based clinical tests are necessarily accompanied. In this paper, we present an objective method to evaluate and predict the performance of hearing aids without the help of such subject-based tests. In the hearing impairment simulation (HIS) algorithm, a sensorineural hearing impairment model is established from auditory test data of the impaired subject being simulated. Also, in the hearing impairment simulation system the abnormal loudness relationships created by recruitment was transposed to the normal dynamic span of hearing. The nonlinear behavior of the loudness recruitment is defined using hearing loss unctions generated from the measurements. The recruitment simulation is validated by an experiment with two impaired listeners, who compared processed speech in the normal ear with unprocessed speech in the impaired ear. To assess the performance, the HIS algorithm was implemented in real-time using a floating-point DSP.

• Journal of radiological science and technology
• /
• v.43 no.5
• /
• pp.397-404
• /
• 2020

Coronavirus disease(COVID-19) is highly infectious disease that directly affects the lungs. To observe the clinical findings from these lungs, the Chest Radiography(CXR) can be used in a fast manner. However, the diagnostic performance via CXR needs to be improved, since the identifying these findings are highly time-consuming and prone to human error. Therefore, Artificial Intelligence(AI) based tool may be useful to aid the diagnosis of COVID-19 via CXR. In this study, we explored various Deep learning(DL) approach to classify COVID-19, other viral pneumonia and normal. For the original dataset and lung-segmented dataset, the pre-trained AlexNet, SqueezeNet, ResNet18, DenseNet201 were transfer-trained and validated for 3 class - COVID-19, viral pneumonia, normal. In the results, AlexNet showed the highest mean accuracy of 99.15±2.69% and fastest training time of 1.61±0.56 min among 4 pre-trained neural networks. In this study, we demonstrated the performance of 4 pre-trained neural networks in COVID-19 diagnosis with CXR images. Further, we plotted the class activation map(CAM) of each network and demonstrated that the lung-segmentation pre-processing improve the performance of COVID-19 classifier with CXR images by excluding background features.

• Elastomers and Composites
• /
• v.48 no.2
• /
• pp.114-124
• /
• 2013

To study the effects of structural difference and fatty acid chain length of glyceride, new dispersion agents having various glyceride structures such as mono-, di-, and tri-, were prepared using glycerol extracted from palm oil and fatty acid having various chain length ranges from 12 to 18. These dispersion agents were mixed with the rubber compounds and compared with conventional metal salt dispersion agents. Glyceride dispersion agent provided remarkable improvement in silica dispersion, compared to metal salt fatty acidic one, even though the viscosity of mixtures was relatively high due to low lubricating effect, and this was approved by mechanical properties, wear properties, and Payne effect. Also, the longer in chain length of fatty acid and the smaller in numbers of fatty acid, the dispersity of silica was improved.

• Aerospace Engineering and Technology
• /
• v.5 no.2
• /
• pp.16-26
• /
• 2006

GNSS is a satellite-based radio navigation aid. For using it in civil air navigation area, any GNSS service should meet the requirements of accuracy, integrity, continuity and availability in each flight phase established by ICAO. In this study, a remote integrity monitoring system(RIMS) for GNSS are proposed and explained to utilize it in the design of GNSS augmentation system such as GBAS and GRAS. The RIMS consists of signal-in-space receiving subsystem and signal processing subsystem. Each GPS receiver is connected to Host PC by the serial to ethernet converting device which is able to convert serial port connection to LAN port connection in order to exchange information via the internet. We can overcome the siting limitation of GPS receiver and antenna, and reduce signal loses in the cable between GPS antenna and receiver. This system is providing the development environment for GBAS CAT-I system.

• Phonetics and Speech Sciences
• /
• v.12 no.3
• /
• pp.55-63
• /
• 2020

Voice conversion can be applied to various voice processing applications. It can also play an important role in data augmentation for speech recognition. The conventional method uses the architecture of voice conversion with speech synthesis, with Mel filter bank as the main parameter. Mel filter bank is well-suited for quick computation of neural networks but cannot be converted into a high-quality waveform without the aid of a vocoder. Further, it is not effective in terms of obtaining data for speech recognition. In this paper, we focus on performing voice-to-voice conversion using only the raw spectrum. We propose a deep learning model based on the transformer network, which quickly learns the voice conversion properties using an attention mechanism between source and target spectral components. The experiments were performed on TIDIGITS data, a series of numbers spoken by an English speaker. The conversion voices were evaluated for naturalness and similarity using mean opinion score (MOS) obtained from 30 participants. Our final results yielded 3.52±0.22 for naturalness and 3.89±0.19 for similarity.

• The KIPS Transactions:PartD
• /
• v.9D no.1
• /
• pp.167-172
• /
• 2002

Recently, some of people are keep in research and development of the further more efficient and convenient auction systems using intelligent software agents in electronic commerce. The purpose of this thesis is that a simple auction system has web bulletin boards, is aided by intelligent agent, and generates pertinent auction duration time and starting price for auction goods of auctioneer into a auction system, then the auctioneer gets the highest margin. The seller who want to sell goods, is using internet sends mail that has information for goods to agent of internet auction system. The agent undertake filtering process for already learned information about similar goods. And it calculate duration time and start price from stored bidding history database. In this thesis we propose a mailing agent system pushing information in internet auction that enables to aid decision for auctioneer about the starting time and price which delivers the highest margin.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.42 no.12
• /
• pp.165-170
• /
• 2005

In this paper, influences of a ultra wideband (UWB) antenna on impulse channel measurement are investigated in time domain (TD) and frequency domain (FD) as well. Firstly, impulse response of an UWB antenna is obtained and then using the result of impulse response of the UWB antenna, influences of the antenna on impulse radio channel is analyzed. Furthermore, using the impulse response of the UWB anenna, method of impulse radio channel analysis is presented by excluding the effect of the antenna from an impulse radio channel. For verifying the theory, a modified conical monopole antenna is designed for measuring impulse radio channel and its impulse response is obtained. After that, in order to investigate the effects of the UWB antenna on an impulse radio channel, multipath environments are set up in an anechonic chamber and transmission coefficient for each multipath environment is measured with an aid of vector network analyzer. Data measured in frequency domain is transformed into those in time domain by way of signal processing. Measurement shows that such properties of the antenna as dispersion and ringing affect impulse radio channel. Moreover, using the impulse response of the antenna, impulse response of only multipath channel is obtained.

• Korean Journal of Agricultural Science
• /
• v.32 no.1
• /
• pp.81-94
• /
• 2005

This study is aimed at giving the basic information for individual farm households to make decisions for optimizing their farm sizes and for the government to implement farm size optimization policies through the identification of combinations among rice production factors in plain areas like Cheolwon district and the suggestion of the optimal farm sizes of individual farmers based on the scale of economy calculated. The data of agricultural production costs of 50 rice farmers in the plain area which is located in Dongsong-eup Cholwon district, Kangwon province were used in the analysis. The 'translog' cost function among various methods which is a flexible function type was adopted to calculate the scale of economy in rice production. Seemingly unrelated regression(SUR) method was used in forecasting functions and processing other statistics by SHAZAM which is one of the computer aid program for quantitative econometric analysis. In conclusion, the long-run average cost(LAC) curve showed 'U-shape' which was different from 'L-type' one which was shown in the previous studies by others. The lowest point of the LAC was 9.764ha and the concerned production cost amounted to 633 Won/kg. Based on these results, it have to be suggested that around 10 ha of paddy is the target size for policy assistances to save costs under the present level of farming practices and technology. The above results show that the rice production costs could be saved up to 10ha in Cheolwon plain area which is a typical paddy field. However, land use, land condition, land ownership and manager's ability which may affect scale of economy should be considered. Furthermore, reasonable management will have to be realized by means of labor saving technology and cost saving management skill like enlargement of farm size of rice.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.12
• /
• pp.1772-1781
• /
• 2017

In this study, the design methodology for alleviating the overfitting problem of Polynomial Neural Networks(PNN) is realized with the aid of two kinds techniques such as L2 regularization and Sum of Squared Coefficients (SSC). The PNN is widely used as a kind of mathematical modeling methods such as the identification of linear system by input/output data and the regression analysis modeling method for prediction problem. PNN is an algorithm that obtains preferred network structure by generating consecutive layers as well as nodes by using a multivariate polynomial subexpression. It has much fewer nodes and more flexible adaptability than existing neural network algorithms. However, such algorithms lead to overfitting problems due to noise sensitivity as well as excessive trainning while generation of successive network layers. To alleviate such overfitting problem and also effectively design its ensuing deep network structure, two techniques are introduced. That is we use the two techniques of both SSC(Sum of Squared Coefficients) and $L_2$ regularization for consecutive generation of each layer's nodes as well as each layer in order to construct the deep PNN structure. The technique of $L_2$ regularization is used for the minimum coefficient estimation by adding penalty term to cost function. $L_2$ regularization is a kind of representative methods of reducing the influence of noise by flattening the solution space and also lessening coefficient size. The technique for the SSC is implemented for the minimization of Sum of Squared Coefficients of polynomial instead of using the square of errors. In the sequel, the overfitting problem of the deep PNN structure is stabilized by the proposed method. This study leads to the possibility of deep network structure design as well as big data processing and also the superiority of the network performance through experiments is shown.