• Journal of radiological science and technology
• /
• v.46 no.4
• /
• pp.277-285
• /
• 2023

In this study, pneumonia identification networks with the small number of layers were constructed by using chest X-ray images. The networks had similar trainable-parameters, and the performance of the trained models was quantitatively evaluated with the modification of the network architectures. A total of 6 networks were constructed: convolutional neural network (CNN), VGGNet, GoogleNet, residual network with identity blocks, ResNet with bottleneck blocks and ResNet with identity and bottleneck blocks. Trainable parameters for the 6 networks were set in a range of 273,921-294,817 by adjusting the output channels of convolution layers. The network training was implemented with binary cross entropy (BCE) loss function, sigmoid activation function, adaptive moment estimation (Adam) optimizer and 100 epochs. The performance of the trained models was evaluated in terms of training time, accuracy, precision, recall, specificity and F1-score. The results showed that the trained models with the small number of layers precisely detect pneumonia from chest X-ray images. In particular, the overall quantitative performance of the trained models based on the ResNets was above 0.9, and the performance levels were similar or superior to those based on the CNN, VGGNet and GoogleNet. Also, the residual blocks affected the performance of the trained models based on the ResNets. Therefore, in this study, we demonstrated that the object detection networks with the small number of layers are suitable for detecting pneumonia using chest X-ray images. And, the trained models based on the ResNets can be optimized by applying appropriate residual-blocks.

• Korean Journal of Food Science and Technology
• /
• v.42 no.4
• /
• pp.390-397
• /
• 2010

Residual solvents in foods are defined as organic volatile chemicals used or produced in manufacturing of extracts or additives, or functional foods. The solvents are not completely eliminated by practical manufacturing techniques and they also may become contaminated by solvents from packing, transportation or storage in warehouses. Because residual solvents have no nutritional value but may be hazardous to human health, there is a need to remove them from the final products or reduce their amounts to below acceptable levels. The purpose of this study was to develop and evaluate an analytical method for the screening of residual solvents in health functional foods. Furthermore, the aim of this study was to constitute a reasonable management system based on the current state of the market and case studies of foreign countries. Eleven volatile solvents such as MeOH, EtOH, trichloroethylene and hexane were separated depending on their column properties, temp. and time using Gas Chromatography (GC). After determining the GC conditions, a sample preparation method using HSS (Head Space Sampling) was developed. From the results, a method for analyzing residual solvents in health functional foods was developed considering matrix effect and interference from the sample obtained from the solution of solvents-free health functional foods spiked with 11 standards solutions. Validation test using the developed GC/HSS/MS (Mass Spectrometry) method was followed by tests for precision, accuracy, recovery, linearity and adequate sensitivity. Finally, examination of 104 samples grouped in suits was performed by the developed HSS/GC/MS for screening the solvents. The 11 solvents were isolated from health functional foods based on vapor pressure difference, and followed by separation within 15 minutes in a single run. The limt of detection (LOD), limit of quantification (LOQ), recovery and coefficient of variation (C.V.) of these compounds determined by the HSS/GC/MS were found to be 0.1 pg/mL, 0.1-125 pg/g, 51.0-104.6%, and less than 15%, respectively. Using the developed HSS/GC/MS method, residual solvent from 16 out of 104 health functional products were detected as a EtOH. This method therefore seems t o be a valuable extension ofanalytical method for the identification of residual solvents in health functional food.

• Journal of Korea Foundry Society
• /
• v.15 no.2
• /
• pp.175-183
• /
• 1995

It is difficult to remove such interstitial impurities as sulfur, oxygen, hydrogen and carbon in Fe-Ni alloys. Thermodynamic and kinetic studies were carried out on the behavior of hydrogen gas, oxygen gas, Si, Al and slag, and the reaction time by the $Ar/Ar-H_2$ plasma and electron beam melting. After the addition of Al, Si, they were melted by Ar plasma with reaction time changed. 80%Ni-Fe alloys showed a better deoxidation than 36%Ni-Fe alloys. At $Ar-H_2$ plasma melting, the deoxidation was significant. In the case of the electron beam melting, the residual oxygen was higher than in Ar plasma melting because electron beam melting temperature was lower than that of Ar plasma. For the decaburization, it was melted by $Ar-O_2$ plasma melting, which could remove effectively carbon by activated oxygen in plasma. We added slag to Fe-Ni alloys for the desulfurization. As the result of this experiments, the amount of residual sulfur was not changed according to the slag ratio and reaction time.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.4 no.3
• /
• pp.190-196
• /
• 2009

A new time adapted threshold using the standard deviations of Wavelet coefficients after Wavelet transform by frame scale is proposed. The time adapted threshold is set up using the sum of standard deviations of Wavelet coefficient in level 3 approximation and weighted level 1 detail. Level 3 approximation coefficients represent the voiced sound with low frequency and level 1 detail coefficients represent the unvoiced sound with high frequency. After reducing noise by soft thresholding with the proposed time adapted threshold, there are still residual noises in silent interval. To reduce residual noises in silent interval, a detection algorithm of silent interval is proposed. From simulation results, it can be noticed that SNR and MSE of the proposed algorithm are improved than those of Wavelet transform and than those of Wavelet packet transform.

• Structural Engineering and Mechanics
• /
• v.73 no.3
• /
• pp.303-317
• /
• 2020

In this study an innovative rocking zipper braced frame (RZBF) is proposed to overcome the deficiencies of common concentrically braced frames. RZBF is an improved rocking concentrically braced frame which is based on combination of rocking behavior and zipper columns. The base rocking joints and post-tensioned bars provide rocking response and restoring force, respectively. Also, zipper columns distribute the unbalance force over the frame height and reduce the damage concentration. To evaluate seismic performance of RZBF, a comparison study is carried out considering concentrically braced frame, zipper braced frame, rocking concentrically braced frame and RZBF. Thereby, a suite of non-linear time history analyses had been performed on four different types of archetypes with four, six, eight, ten and twelve stories. Frames were designed and non-linear time history analyses were conducted in OpenSees. To compare the seismic behavior of the archetypes, roof drifts, residual roof drifts, story drifts, the forces of first and top story braces, PT bars forces, column uplift and base shears were taken in to consideration. Results illustrate that using RZBF, can reduce the damage due to reduced residual drifts. Zipper columns enhance the seismic performance of rocking systems. As the number of stories increase in the RZBF systems, larger top story braces were needed. So the RZBF system is applicable on low and midrise buildings.

• Journal of Internet Computing and Services
• /
• v.10 no.3
• /
• pp.61-69
• /
• 2009

Most existing clustering protocols have been aimed to provide balancing the residual energy of each node and maximizing life-time of wireless sensor networks. In this paper, we present the threshold based cluster head replacement strategy for clustering protocols in wireless sensor networks. This protocol minimizes the number of cluster head selection by preventing the cluster head replacement up to the threshold of residual energy. Reducing the amount of head selection and replacement cost, the life-time of the entire networks can be extended compared with the existing clustering protocols. Our simulation results show that our protocol outperformed than LEACH in terms of balancing energy consumption and network life-time.

• Korean Journal of Veterinary Research
• /
• v.43 no.3
• /
• pp.435-438
• /
• 2003

The objective of this study was to investigate the residue depletion of cephaalexin in the flounder (Paralichthys olivaceus) after multiple oral administrations and to establish the appropriate withdrawal time for edible tissues. A highly sensitive and specific method for the determination of cephalexin in the serum of flounder by LC/MS was developed and validated. Mean recoveries from serum were 87.2% (ranged from 81.2% to 94.5%) for cepalexin. Recovery and precision met the criteria for the guideline of residual analysis of veterinary drugs by the National Veterinary Research and Quarantine Service (NVRQS) in Korea. The limit of detection and limit of quantitation of cephalexin were 10 ng/ml and 50 ng/ml, respectively. Residual levels of cephalexin in muscle samples were estimated with 95% tolerance limit and 95% confidence to fall below the MRL after a withdrawal time of 4 days and 5 days for the 40 and 160 mg/kg/day, respectively.

• Journal of the Korea Computer Industry Society
• /
• v.6 no.2
• /
• pp.271-278
• /
• 2005

This paper proposed a new time adapted threshold using the standard deviations of Wavelet coefficients after Wavelet transform by frame scale. The time adapted threshold is set up using the sum of standard deviations of Wavelet coefficient in level 3 approximation and weighted level 1 detail. Level 3 approximation coefficients represent the voiced sound with low frequency and level 1 detail coefficients represent the unvoiced sound with high frequency. After reducing noise by soft thresholding with the proposed time adapted threshold, there are still residual noises in silent interval. To reduce residual noises in silent interval, a detection algorithm of silent interval is proposed. From simulation results, it is demonstrated that the proposed algorithm improves SNR and MSE performance more than Wavelet transform and Wavelet packet transform does.

• The Journal of Korean Medicine
• /
• v.21 no.4
• /
• pp.236-241
• /
• 2000

Background and Purpose : Cerebrovascular accident is a serious neurologic event. It can have temporary or permanent effects on survivors, including memory, cognition and volitional control of voiding. The purpose of this study is to evaluate the clinical effect of moxibustion in patients with voiding dysfunction after a cerebrovascular accident. Methods : Twenty patients with post-stroke voiding dysfunction were studied. All patients had computerized tomography(CT) and magnetic resonance imaging(MRI) to localize the lesion in the central nervous system. They were randomly divided into two groups : the control and moxibustion group. Ten of twenty patients underwent moxibustion treatment by randomization. The moxibustion group receieved moxibustion at three points : Chung-guk (Conception Vessel CV3), Kuanwon (CV4) Kihae (CV6). Residual urine volume evaluation was undertaken in all patients. Results : 1. IIn the balanced bladder time, the moxibustion group had the shorter time than non moxibustion group. But there was no significiant difference between the moxibustion group and non-moxibustion group. 2. In the residual urine volume, the moxibustion group and non-moxibustion group showed a tendency to decrease. But there was no significiant difference between the moxibustion group and non-moxibustion group. Conclusions : Though further study is needed, our findings suggested that the time of achieving balanced voiding was shorter with moxibustion than in the control group.

• Structural Engineering and Mechanics
• /
• v.63 no.4
• /
• pp.497-507
• /
• 2017

During recent earthquakes, a significant number of concrete structures suffered extensive damage. Conventional reinforced concrete structures are designed for life-time safety that may see permanent inelastic deformation after severe earthquakes. Hence, there is a need to utilize adequate materials that have the ability to tolerate large deformation and get back to their original shape. Super-elastic shape memory alloy (SMA) is a smart material with unique properties, such as the ability to regain undeformed shape by unloading or heating. In this research, four different stories (three, five, seven and nine) of reinforced concrete (RC) buildings have been studied and subjected to near-field ground motions. For each building, two different types of reinforcement detailing are considered, including (1) conventional steel reinforcement (RC frame) and (2) steel-SMA reinforcement (SMA RC frame), with SMA bars being used at plastic zones of beams and steel bars in other regions. Nonlinear time history analyses have been performed by "SeismoStruct" finite element software. The results indicate that the application of SMA materials in plastic hinge regions of the beams lead to reduction of the residual displacement and consequently post-earthquake repairs. In general, it can be said that shape memory alloy materials reduce structural damage and retrofit costs.