• 대한의용생체공학회:의공학회지
• /
• 제39권6호
• /
• pp.229-236
• /
• 2018

Liver cancer is one of the highest incidents in the world, and the mortality rate is the second most common disease after lung cancer. The purpose of this study is to evaluate the diagnostic ability of deep learning in the classification of malignant and benign tumors in CT images of patients with liver tumors. We also tried to identify the best data processing methods and deep learning models for classifying malignant and benign tumors in the liver. In this study, CT data were collected from 92 patients (benign liver tumors: 44, malignant liver tumors: 48) at the Gil Medical Center. The CT data of each patient were used for cross-sectional images of 3,024 liver tumors. In AlexNet and VggNet, the average of the overall accuracy at each image size was calculated: the average of the overall accuracy of the $200{\times}200$ image size is 69.58% (AlexNet), 69.4% (VggNet), $150{\times}150$ image size is 71.54%, 67%, $100{\times}100$ image size is 68.79%, 66.2%. In conclusion, the overall accuracy of each does not exceed 80%, so it does not have a high level of accuracy. In addition, the average accuracy in benign was 90.3% and the accuracy in malignant was 46.2%, which is a significant difference between benign and malignant. Also, the time it takes for AlexNet to learn is about 1.6 times faster than VggNet but statistically no different (p > 0.05). Since both models are less than 90% of the overall accuracy, more research and development are needed, such as learning the liver tumor data using a new model, or the process of pre-processing the data images in other methods. In the future, it will be useful to use specialists for image reading using deep learning.

• 한국연소학회:학술대회논문집
• /
• 한국연소학회 2006년도 제33회 KOSCO SYMPOSIUM 논문집
• /
• pp.97-103
• /
• 2006

This work forcus on the development of gas sensor that measure the concentrations of exhaust gas using diode laser. Each diode laser for exhaust gas measurement is set to work at near-IR using both DA and WMS methods. Also use of fiber-coupled optical elements makes such a sensor rugged and easy to align. On-line data acquisition and processing can be performed with a PC running LabVIEW software, and absorption signals are measured simultaneously by multiplexing method. Finally, It were experimentally compared WMS (Wavelength Modulation Spectroscopy) with DA (Direct Absorption) for the accuracy.

• 전자공학회논문지D
• /
• 제34D권8호
• /
• pp.15-25
• /
• 1997

The integration method of carier concentrations to redcue the discretization error of th box integratio method used in the discretization of the three-dimensional poisson's equation is presented. The carrier concentration is approximated in the closed form as an exponential function of the linearly varying potential in the element. The presented method is implemented in the three-dimensional poisson's equation solver running under the windows 95. The accuracy and the convergence chaacteristics of the three-dimensional poisson's equation solver are compared with those of DAVINCI for the PN junction diode and the n-MOSFET under the thermal equilibrium and the DC reverse bias. The potential distributions of the simulatied devices from the three-dimensional poisson's equation solver, compared with those of DAVINCI, has a relative error within 2.8%. The average number of iterations needed to obtain the solution of the PN junction diode and the n-MOSFET using the presented method are 11.47 and 11.16 while the those of DAVINCI are 21.73 and 23.0 respectively.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2005년도 추계종합학술대회
• /
• pp.1083-1086
• /
• 2005

Modern superscalar processors exploit Instruction Level Parallelism to achieve high performance by speculative techniques such as branch prediction. The indirect branch target prediction is very difficult compared to the prediction of direct branch target and branch direction, since it has dynamically polymorphic target. We present a accurate and hardware-efficient indirect branch target predictor. It can reduce the tags which has to be stored in the Indirect Branch Target Cache without a sacrifice of the prediction accuracy. We implement the proposed scheme on SimpleScalar and show the efficiency running SPEC95 benchmarks.

• 전기학회논문지
• /
• 제56권9호
• /
• pp.1688-1693
• /
• 2007

The aim of this paper is to develop and verify the system which can detect heart rate during exercise by using conductive fabric electrode and transportable measurement module. The experiment was performed under 4 conditions(resting, walking, jogging, running) and 18 subjects data are used. By using the ECG measurement system used in cardiac stress testing as reference value in order to verify the accuracy of the developed system, the relative error and correlation coefficient was calculated for each subject at every 3 seconds. The results have shown that the high correlation between the developed system and the reference system for detecting heart rate during exercise. Relative error and correlation coefficient are 2.27% and 0.9877, respectively. 7 subjects data are omitted in these calculations because of severe noises. Therefore, it is expected that this system could be used as a health monitoring system in ubiquitous environment in the future.

• 산업경영시스템학회지
• /
• 제16권27호
• /
• pp.121-125
• /
• 1993

This study deals with the estimation of the Weibull parameters, which have a close relation with product reliability characteristics. Among the many kinds of estimation methods, Kao's Weibull Probability Paper(WPP) is commonly used. The WPP is very convenient, but it has a great disadvantage in estimation accuracy by plotting method. It is very difficult to get the same results even if one use the same data several times. A computer program for the regression method is used for the parameter estimation to reduce these errors. Especially, the computer graphic program was written in GW-BASIC 3.22 language and the program appears in the appendix part with a couple of running examples for user's reference.

• 한국자동차공학회논문집
• /
• 제3권3호
• /
• pp.97-108
• /
• 1995

A 3-dimensional double track vehicle model, that has 12-degress-of-freedom, was proposed to analyze handling and riding behaviours of an automotive car. Nonlinear characteristics of the suspension and steering systems of the vehicle model were considered in its equations of motion, which were solved by using the 4th-order Runge-Kutta integration method. Computer simulations for lane change, steady-state handling, and running-over-bump manoeuvres were made and verified by vehicle tests on proving ground. The computed results of the proposed model showed better agreement with test results than those of the conventional 2-dimensional single track model did. Especially they showed good accuracy near the characteristic speed and in high lateral accelerated manoeuvres.

• 소음진동
• /
• 제10권3호
• /
• pp.457-464
• /
• 2000

Vibrations of a fan are often generated by mechanical unbalance magnetic force and air flow. These vibrations depend on the design of a fan the machining accuracy of each element and assembled conditions. An experimental study is carried out to reduce the vibration and noise of an industrial fan in this paper. In order to identify the vibration sources of a fan the signal analysis and system analysis are performed, It is shown that the industrial fan studied in this paper has a natural frequency at 144 Hz and resonance occurs when the running speed of the fan is 1750 rpm. The results may be helpful to design a fan with low vibration and noise.

• 전기학회논문지
• /
• 제58권7호
• /
• pp.1443-1449
• /
• 2009

This paper presents a profiling model of a wide-window superscalar microprocessor using multiple branch prediction. The key idea is to apply statistical profiling technique to the superscalar microprocessor with a wide instruction window and a multiple branch predictor. The statistical profiling data are used to obtain a synthetical instruction trace, and the consecutive multiple branch prediction rates are utilized for running trace-driven simulation on the synthesized instruction trace. We describe our design and evaluate it with the SPEC 2000 integer benchmarks. Our performance model can achieve accuracy of 8.5 % on the average.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2017년도 추계학술대회
• /
• pp.47-48
• /
• 2017

This paper presents a low-cost portable lithium battery parameter measuring and estimating the solution. In this method, lithium battery characteristics are monitored during discharging and charging cycles. The battery profile is analyzed, and its key parameters are estimated by GNU Octave running on Raspberry Pi Zero, a mini computer. The proposed method can measure and estimate the battery parameters for SOC and DOD estimation with reasonable accuracy as well as portability features.