• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.2
• /
• pp.90-96
• /
• 2023

Securing large amounts of training data in deep learning neural networks, including convolutional neural networks, is of importance for avoiding overfitting phenomenon or for the excellent performance. However, securing labeled training data in deep learning neural networks is very limited in reality. To overcome this, several augmentation methods have been proposed in the literature to generate an additional large amount of training data through transformation or manipulation of the already acquired traing data. However, unlike training data such as images and texts, it is barely to find an augmentation method in the literature that additionally generates bio-signal training data for convolutional neural network based human activity recognition. Thus, this study proposes a simple but effective augmentation method of bio-signal training data for convolutional neural network based human activity recognition. The usefulness of the proposed augmentation method is validated by showing that human activity is recognized with high accuracy by convolutional neural network trained with its augmented bio-signal training data.

• Applied Chemistry for Engineering
• /
• v.23 no.3
• /
• pp.279-283
• /
• 2012

Bio-ethanol production research using various material has been problemed for solving problems of environment pollution caused by fossil fuels. Red-algae consists of agar, carrageenan, and porphyran. If it is treated by acid, it is able to change useful bio-mass for bio-ethanol. In this study, we found an optimal condition for bio-ethanol production from acid hydrolysate in red-algae. To produce bio-ethanol, Saccharomyces cerevisiae KCCM1129 inoculated to acid hydrolysate of Gelidium amansii. The optimal condition for Gelidium amansii hydrolysis was found to be 30 min reaction at $H_2SO_4$ concentration of 1.5% and $121^{\circ}C$. At this condition, its produced to 7.04 g/L galactose and 1.94 g/L glucose. And acetic acid concentration of 2.0% in agar produced 0.75 g/L galactose. In contrast, Pachymeniopis elliptica was treated with $H_2SO_4$concentration of 1.5%, it produced 6.38 g/L galactose. And Pachymeniopis elliptica treated with acetic acid concentration of 2% produced 0.368 g/L galactose. The optimal condition of ethanol production was found to be 96 h reaction at $H_2SO_4$concentration of 1.0% and $30^{\circ}C$, which produced 3.77 g/L ethanol.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.47 no.5
• /
• pp.24-32
• /
• 2010

In this study, we have proposed a pre-processing algorithm newly developed for improving the tracking efficiency of the optical flow method. The developed pre-processing algorithm consists of a median filter, binarization, morphology, canny edge, contour detecting and an approximation method. In order to evaluate whether the optical flow tracking capacity increases, this study applied the pre-processing algorithm to the Lucas-Kanade(LK) optical flow algorithm, and comparatively analyzed its images and tracking results with those of optical flow without the pre-processing algorithm and with the existing pre-processing algorithm(composed of median filter and histogram equalization). As a result, it was observed that the tracking performance derived from the LK optical flow algorithm with the pre-processing algorithm, shows better tracking accuracy, compared to the one without the pre-processing algorithm and the one with the existing pre-processing algorithm. It seems to have resulted by successful segmentation for characteristic areas and subdivision into inner and outer contour lines.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.14 no.5
• /
• pp.413-420
• /
• 2021

This study is about a machine learning model for predicting insufficient parameters through other parameters when one of the collected parameters is insufficient. A regression model was created to predict time, temperature, humidity, CO2, and light quantity data through the machine learning regression analysis function in Matlab. In addition, the three models with the lowest RMSE values for each parameter were selected and verified. For verification, the predicted values were obtained by applying the test data to the prediction model derived from each parameter, and the correlation coefficient and error average between the measured values and the obtained predicted values were obtained and then compared.

• Composites Research
• /
• v.27 no.5
• /
• pp.196-200
• /
• 2014

A novel, simple and totally recyclable method has been developed for the synthesis of nontoxic, biocompatible and biodegradable bio-composite films from soy protein and silk protein. Bio films are defined as flexible films prepared from biological materials such as protein. These materials have potential application in medical and food as a packaging material. Their use depends on various parameters such as mechanical (strength and modulus), thermal, among others. In this study, prepare and characterization of bio films made from Soy Protein Isolate (SPI) (matrix) and Silk Fiber (SF) (reinforcement) through solution casting method by the addition of plasticizer and crosslinking agent. The obtained SPI and SPI/SF composites were subsequently subjected to evaluate their mechanical and thermal properties by using Universal Testing Machine and Thermal Gravimetric Analyzer respectively. The tensile testing showed significant improvements in strength with increasing amount of SF content and the % elongation at break of the composites of the SPI/SF was lower than that of the matrix. Though the interfacial bonding was moderate, the improvement in tensile strength and modulus was attributed to the higher tensile properties of the silk fiber.

• Journal of Biomedical Engineering Research
• /
• v.32 no.1
• /
• pp.37-44
• /
• 2011

This paper describes the development and assessment of conductive fabric sensor for evaluating knee movement using bio-impedance measurement method. The proposed strip-typed conductive fabric sensor is compared with a dot-typed Ag/AgCl electrode for evaluating validity under knee movement condition. Subjects are composed of ten males($26.6{\pm}2.591$) who have not had problems on their knee. The strip-typed conductive fabric sensor is analyzed by correlation and reliability between a dot-typed Ag/AgCl electrode and the strip-typed conductive fabric sensor. The difference of bio-impedance between a dot-typed Ag/AgCl electrode and the strip-typed conductive fabric sensor averages $7.067{\pm}13.987\;{\Omega}$ As the p-value is under 0.0001 in 99% of t-distribution, the strip-typed conductive fabric sensor is correlated with a dot-typed Ag/AgCl electrode by SPSS software. The strip-typed conductive fabric sensor has reliability when it is compared with a dot-typed Ag/AgCl electrode because most of bio-impedance values are in ${\pm}1.96$ standard deviation by Bland-Altman Analysis. As a result, the strip-typed conductive fabric sensor can be used for assessing knee movement through bio-impedance measurement method as a dot-typed Ag/AgCl electrode. Futhermore, the strip-typed conductive fabric sensor is available for wearable circumstances, applications and industries in the near future.

• Progress in Medical Physics
• /
• v.25 no.2
• /
• pp.116-121
• /
• 2014

Next future, The bio technology will be a rapidly developing. This paper is the scattering beam measurement of the red blood cell (RBC) and the fabrication of the micro cell biochip using the bio micro electro mechanical system (Bio-MEMS) process technology. The Major process method of Bio-MEMS technology was used the buffered oxide etchant (BOE), electro chemical discharge (ECD) and ultraviolet sensitive adhesives (UVSA). All experiments were the 10 times according to the process conditions. The experiment and research are required the ultraviolet expose, the micro fluid current, the cell control and the measurement of the output voltage Vpp (peak to peak) waveform by scattering angles. The transmitting and receiving of the laser beam was used the single mode optical fiber. The principles of the optical properties are as follows. The red blood cells were injected into the micro channel. The single mode optical fiber was inserting in the guide channel. The He-Ne laser beam was focusing in the single mode optical fiber. The transmission He-Ne laser beam is irradiating to the red blood cells. The manufactured guide channel consists of the four inputs and the four outputs. The red blood cell was allowed with the cylinder pump. The output voltage Vpp waveform of the scattering beam was measured with a photo detector. The receiving angle of the output optical fiber is $0^{\circ}$, $5^{\circ}$, $10^{\circ}$, $15^{\circ}$. The magnitude of the output voltage Vpp waveform was measured in the decrease according to increase of the reception angles. The difference of the output voltage Vpp waveform is due differences of the light transmittance of the red blood cells.

• Journal of the Korean Society of Safety
• /
• v.35 no.1
• /
• pp.62-69
• /
• 2020

Most people suffer from stress during day life because modernized society is very complex and changes fast. Because stress can affect to many kind of physiological phenomena it is even considered as a disease. Therefore, it should be detected earlier, then must be released. When a person is being stressed several bio-signals such as heart rate, etc. are changed. So, those can be detected using medical electronics techniques. In this paper, stress assessment system is studied using random forest algorithm based on heart rate, RR interval and Galvanic skin response. The random forest model was trained and tested using the data set obtained from the bio-signals. It is found that the stress assessment procedure developed in this paper is very useful.

• Journal of Korea Multimedia Society
• /
• v.19 no.8
• /
• pp.1236-1247
• /
• 2016

Recently, the level set has become a popular method in many research fields. The main reason is that it can be modified into many variants. One such case is our proposed method. We describe a contrast-enhancement method to segment the hippocampal region from the background. However, the hippocampus region has quite similar intensities to the neighboring pixel intensities. In addition, to handle the inhomogeneous intensities of the hippocampus, we used a bias correction before hippocampal segmentation. Thus, we developed a contrast-enhanced bias-corrected distance-regularized level set (CBDLS) to segment the hippocampus in magnetic resonance imaging (MRI). It shows better performance than the distance-regularized level set evolution (DLS) and bias-corrected distance-regularized level set (BDLS) methods in 33 MRI images of one normal patient. Segmentation after contrast enhancement and bias correction can be done more accurately than segmentation while not using a bias-correction method and without contrast enhancement.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.1075-1076
• /
• 2008

Hemorrhagic shock is a common cause of death in emergency rooms. The objective evaluation of hemorrhagic shock is very important for early diagnosis and treatment. The purpose of this study is to understand its mechanism by analyzing the changes of bio-signals in hemorrhagic shock using controlled hemorrhage of SD rats. In this study, we constructed a hemorrhagic integrated system to control bleeding and to simultaneously measure bio-signals such as ECG, blood pressure, temperature, and respiration. In order to verify the system, we measured the bio-signals mentioned above while hemorrhagic shock was induced by withdrawing blood (2.5ml/100g/15min) from a femoral vein for 10 rats.