Search

In past decades, stretchable conductors have been investigated for a wide range of applications, and the operation strain range of such devices varies by application. To commercialize stretchable devices, it is necessary to optimize the deformation of stretchable electrodes based on a given device elongation range. Therefore, we investigated the deformation mechanics of a silver nanowire (AgNW) electrode on an elastomeric substrate depending on its junction treatment method. At low-strain ( 20%), the thermally annealed AgNW electrode rapidly increased in resistance, while the laser-welded AgNW electrode showed lower resistivity change. By in situ surface analysis and a repetitive tensile test, we observed that the thermally annealed AgNW electrode shows less cracking at low strain but the laser-welded AgNW electrode exhibits fracturing of individual nanowires at low strain. Furthermore, at high strain, laser-welded AgNWs could slide to reduce stress during elongation, resulting in a smaller change in resistance compared to that of thermally annealed AgNW electrode. These results indicate that optimization of adhesion is necessary to fabricate stretchable devices based on deformation range.

Due to the risk of data leakage while training deep learning models in a shared environment, we propose a new privacy-preserving deep learning (PPDL) method using a structural image de-identification approach for object classification. The proposed structural image de-identification approach is designed based on the fact that the degree of structural distortion of an image object has the greatest impact on human's perceptual system. Thus, by modifying only the structural parts of the original one using order preserving encryption(OPE), the proposed structural image de-identification approach decreases only the recognition rate by human. From the experimental results using different standard datasets, we show that the object classification accuracy of the proposed structural image de-identification method is almost the same as the deep learning performance for non-encrypted images, without revealing the original image contents including sensitive information. Also, by handling the trade-off between object classification accuracy and privacy protection for the de-identified image, we experimentally find the optimal size of input image for the proposed structural image de-identification approach.

In this paper, we propose a new deep learning-based image translation method to predict and generate images after hair transplant surgery from images before hair transplant surgery. Since existing image translation models use a naive strategy that trains the whole distribution of translation, the image translation models using the original image as the input data result in converting not only the hair transplant surgery region, which is the region of interest (ROI) for image translation, but also the other image regions, which are not the ROI. To solve this problem, we proposed a novel generative adversarial network (GAN)-based ROI image translation method, which converts only the ROI and retains the image for the non-ROI. Specifically, by performing image translation and image segmentation independently, the proposed method generates predictive images from the distribution of images after hair transplant surgery and specifies the ROI to be used for generated images. In addition, by applying the ensemble method to image segmentation, we propose a more robust method through complementing the shortages of various image segmentation models. From the experimental results using a real medical image dataset, e.g., 1394 images before hair transplantation and 896 images after hair transplantation, to train the GAN model, we show that the proposed GAN-based ROI image translation method performed better than the other GAN-based image translation methods, e.g., by 23% in SSIM (Structural Similarity Index Measure), 452% in IoU (Intersection over Union), and 42% in FID (Frechet Inception Distance), on average. Furthermore, the ensemble method that we propose not only improves ROI detection performance but also shows consistent performances in generating better predictive images from preoperative images taken from diverse angles.

As the amount of data collected and analyzed by machine learning technology increases, data that can identify individuals is also being collected in large quantities. In particular, as deep learning technology—which requires a large amount of analysis data—is activated in various service fields, the possibility of exposing sensitive information of users increases, and the user privacy problem is growing more than ever. As a solution to this user’s data privacy problem, homomorphic encryption technology, which is an encryption technology that supports arithmetic operations using encrypted data, has been applied to various field including finance and health care in recent years. If so, is it possible to use the deep learning service while preserving the data privacy of users by using the data to which homomorphic encryption is applied? In this paper, we propose three attack methods to infringe user’s data privacy by exploiting possible security vulnerabilities in the process of using homomorphic encryption-based deep learning services for the first time. To specify and verify the feasibility of exploiting possible security vulnerabilities, we propose three attacks: (1) an adversarial attack exploiting communication link between client and trusted party; (2) a reconstruction attack using the paired input and output data; and (3) a membership inference attack by malicious insider. In addition, we describe real-world exploit scenarios for financial and medical services. From the experimental evaluation results, we show that the adversarial example and reconstruction attacks are a practical threat to homomorphic encryption-based deep learning models. The adversarial attack decreased average classification accuracy from 0.927 to 0.043, and the reconstruction attack showed average reclassification accuracy of 0.888, respectively.

Flexible micro/nano metal grid transparent conductors emerged as an alternative to the fragile/rigid indium tin oxide electrode. They are usually fabricated by the combination of the conventional photolithography and the vacuum deposition of regular metal grid patterns, however, seriously suffer from moiré and starburst problems induced by periodic regular pattern structures. In this paper, we demonstrated flexible and imperceptible random copper microconductors with an extremely high figure-of-merit (∼2000) by the thermal conduction layer-assisted photonic sintering of copper nanoparticles without damages in the plastic substrate. This process can be easily applied to complicated structures and surfaces including a random pattern which is imperceptible and free of interferences. As a proof-of-concept, a transparent windshield defogger in a car was demonstrated with a Cu transparent random conductor at an extreme and reversible fogging state.

As the representative ensemble machine learning method, the Random Forest (RF) algorithm has widely been used in diverse applications on behalf of the fast learning speed and the high classification accuracy. Research on RF can be classified into two categories: (1) improving the classification accuracy and (2) decreasing the number of trees in a forest. However, most of papers related to the performance improvement of RF have focused on improving the classification accuracy. Only some papers have focused on reducing the number of trees in a forest. In this paper, we propose a new Covariance-Based Dynamic RF algorithm, called C-DRF. Compared to the previous works, while ensuring the good-enough classification accuracy, the proposed C-DRF algorithm reduces the number of trees. Specifically, by computing the covariance between the number of trees in a forest and F-measure at each iteration, the proposed algorithm determines whether to increase the number of trees composing a forest. To evaluate the performance of the proposed C-DRF algorithm, we compared the learning time, the test time, and the memory usage with the original RF algorithm under the different areas of datasets. Under the same or higher classification accuracy, it is shown that the proposed C-DRF algorithm improves the performance of the original RF algorithm by as much as 58.68% at learning time, 47.91% at test time, and 68.06% in memory usage on average. As a practical application area, we also show that the proposed C-DRF algorithm is more efficient than the state-of-the-art RF algorithms in Network Intrusion Detection (NID) area.

Infectious diseases and pandemics, including the COVID-19 pandemic, have a huge economic impact on cities. However, few studies examine the economic resilience of small-scale regions within cities. Thus, this study derives neighborhoods with high economic resilience in a pandemic situation and reveals their urban characteristics. It evaluates economic resilience by analyzing changes in the amount of credit card payments in the neighborhood and classifying the types of neighborhoods therefrom. The study conducted the ANOVA, Kruskal–Wallis, and post hoc tests to analyze the difference in urban characteristics between neighborhood types. Accordingly, three neighborhood types emerged from the analysis: high-resilient neighborhood, low-resilient neighborhood, and neighborhood that benefited from the pandemic. The high-resilient neighborhood is a low-density residential area where many elderly people live. Neighborhoods that benefited are residential areas mainly located in high-density apartments where many families of parents and children live. The low-resilient neighborhood is an area with many young people and small households, many studio-type small houses, and a high degree of land-use mix.

In bioinformatics, researchers have endeavored to resolve the following two issues: 1) how to increase the efficiency of storage through compression and 2) how to provide confidentiality for the genome sequence data. To resolve two issues, the sequence alignment map, the binary alignment map, the compressed reference-oriented alignment map (CRAM), and the selective retrieval on encrypted and CRAM formats were proposed. However, since these formats are stored in a centralized storage that is managed by the genome testing organizations, the privacy of sensitive genome sequence data is not guaranteed. In this paper, we propose a new compressed reference-oriented alignment map, called decentralized storage and compressed reference-oriented alignment map (D-RAM), which preserves the privacy of genome sequence data using a decentralized storage architecture. The proposed D-RAM format uses the reference-based compression and $bzip2$ compression to use storage space efficiently. In addition, to preserve the privacy of genome sequence data, the proposed decentralized storage architecture is designed to store the private genome sequence data and the public genome sequence data separately. From the experimental results under simulation and real genome sequence data, we show that the D-RAM format saves the size of the genome sequence data than other formats. By analyzing the computational complexity with which the attacker recovers the genome sequence data, we also show the theoretical analysis that explains why the D-RAM format is safer than the other formats.

As malware family classification methods, image-based classification methods have attracted much attention. Especially, due to the fast classification speed and the high classification accuracy, Convolutional Neural Network (CNN)-based malware family classification methods have been studied. However, previous studies on CNN-based classification methods focused only on improving the classification accuracy of malware families. That is, previous studies did not consider the cases that the accuracy of CNN-based malware classification methods can be decreased under the existence of adversarial attacks. In this paper, we analyze the robustness of various CNN-based malware family classification models under adversarial attacks. While adding imperceptible non-random perturbations to the input image, we measured how the accuracy of the CNN-based malware family classification model can be affected. Also, we showed the influence of three significant visualization parameters(i.e., the size of input image, dimension of input image, and conversion color of a special character)on the accuracy variation under adversarial attacks. From the evaluation results using the Microsoft malware dataset, we showed that even the accuracy over 98% of the CNN-based malware family classification method can be decreased to less than 7%.

To realize high-speed intrusion detection by accommodating many regular expression (regex)-based signatures and growing network link capacities, we propose the Service TimE–Aware Load-balancing (STEAL) algorithm. This work is motivated from the observation that utilization of a many-core network intrusion detection system (NIDS) is influenced by unfair computational distribution among many-core NIDS nodes. To avoid such unfair computational distribution, STEAL is designed to dynamically distribute a large volume of traffic among manycore NIDS nodes based on packet service time, which is represented by the deep packet time in many-core NIDS nodes. From experiments, we show that compared to the commonly used loadbalancing algorithm based on arrival rate, STEAL increases the number of received packets (i.e., decreases the number of dropped packets) in many-core NIDS. Specifically, by integrating an open source NIDS (i.e. Bro) with STEAL, we show that even under attack-dominant traffic and with many signatures, STEAL can rapidly improve the performance of many-core NIDS to realize highspeed intrusion detection.

In this paper, an engine clutch engagement control for hybrid electric vehicles (HEVs) with a transmission mounted electric device (TMED) will be dealt with. In general, the engine of the HEVs repeatedly turns on and off in even one driving cycle, and when the engine turns on, the fast synchronization of the engine speed and the motor speed which is equal to the transmission input speed is very important to transfer the engine power to the driving wheel without a shock during the engagement for better fuel economy and drivability feeling. A hybrid starter generator (HSG) at HEVs is connected into the engine by a pulley, and the engine speed can be controlled by the HSG. When the engine starts to turn on, the engine torque is very unstable, and unpredictable and then the control of the engine speed by HSG is very difficult. The suggested method consists of a disturbance observer and a tracking controller which is composed of a feedback control and a feedforward control. The disturbance observer estimates and compensates a model uncertainty and disturbances such as an engine torque, a friction torque, and so on. The feedback and feedforward control are used for the engine speed to follow a target speed. Then the suggested method is very robust to the model uncertainty, and the disturbances, and then the fast engagement of the engine clutch without a shock is possible. For the effectiveness of the suggested method, the experimental results by TMED HEVs are shown.

Along with the steady growth of wired and wireless networks, the various new attacks targeting networks are also constantly emerging and transforming. As a efficient way to cope with various attacks, the Random Forest(RF) algorithm has frequently been used as the core engine of intrusion detection because of the faster learning speed and the higher attack detection accuracy. However, the RF algorithm has to input the number of the tree composing the forest as a parameter. In this paper, we proposed a new algorithm that limit the number of trees composing the forest using the McNemar test. To evaluate the performance of the proposed RF algorithm, we compared learning time, accuracy and memory usage of the proposed algorithm with the original RF algorithm and other algorithm by using the KDDcup99 dataset. Under the same detection accuracy, the proposed RF algorithm improves the performance of the original RF algorithm by as much as 97.76% at learning time, 91.86% at test time, and 99.02% in memory usage on average.

Purpose: The efficacy of stenting for the right-side colonic malignant obstruction is unknown. The purpose of this study was to evaluate the safety and feasibility of self-expandable metallic stent insertion for the right-side colonic malignant obstruction and its clinical benefits. Methods: We retrospectively reviewed clinical data from 460 patients who underwent right hemicolectomy for right-side colon cancer from January 2006 to January 2014 at Korea University Anam Hospital. Twenty four patients who developed malignant obstruction in the right -side colon were identified and analyzed. Results: Self-expandable metallic stent insertion was attempted in 14 patients, and initial technical success was achieved in 13 patients (92.9%). No immediate stent-related complications were reported. Complete relief of obstruction was achieved in all of the 13 patients. Eleven patients who failed stenting underwent emergency operation. All of the 13 patients with stent underwent laparoscopic surgery and only one case was converted to open. Only three patients in the emergency group underwent laparoscopic surgery but one was converted. Operative time and number of retrieved lymph nodes did not differ between the two groups. Postoperative hospital stay (9.8±3.2 in stent group vs. 16.3±10.9 days in emergency group, p=0.082) tended to be shorter in the stent group. Estimated blood loss (38.5±138.7 in stent group vs. 381.8± 411.9 in emergency group, p=0.010) and duration for resuming diet (3.2±2.2 in stent group vs. 6.6±7.0 days in emergency group, p=0.017) were significantly better in the stent group. Conclusion: Self-expandable metallic stent appears to be safe and feasible in the right-side colonic malignant obstruction. It facilitates minimally invasive surgery and may result in better short-term surgical outcome.

With the emergency of Bring Your Own Device (BOYD) and Internet of Things (IoT) in enterprises, there has been an increase in the use of wireless networks through mobile devices and tablet computers. However, as a negative effect, such increase has also led to increased security threat in wireless networks and thus, the importance of wireless network security has increased. As a way to cope with such threat, many enterprises have introduced wireless intrusion prevention system (WIPS) solution, which consists of WIPS sensors and a centralized WIPS server. By monitoring the radio signal in wireless local area networks (LAN), WIPS solution shuts off access by an unauthorized wireless device and prevents an rogue Access Point (AP). In this paper, we propose an implementation framework for WIPS solution that ananlzes the received signal strength (RSS) values collected from multiple WIPS sensors. By overcoming the performance (including accuracy and F-measure) decrease in the wireless environment with much noise, the proposed WIPS solution shows the high performance even under the wireless environment with low signal-to-noise ratio (SNR). For this purpose, the proposed WIPS solution analyzes the RSS values collected from multiple WIPS sensors by using a dynamic random forest (DRF) algorithm. By adding trees that are the most suitable to forest, the DRF algorithm can improve the performance of the WIPS solution. From the experimental results, it is observed that the proposed WIPS solution combined with the DRF algorithm shows the good-enough performance under various SNR and attack distance.

Purpose: A simulator for the design and performance evaluation of a tractor with a hydro-mechanical transmission (HMT) was developed. Methods: The HMT consists of a hydro-static unit (HSU), a swash plate control system, and a planetary gear. It was modeled considering the input/output relationship of the torque and speed, and efficiency of HSU. Furthermore, a dynamic model of a tractor was developed considering the traction force, running resistance, and PTO (power take off) output power, and a tractor performance simulator was developed in the co-simulation environment of AMESim and MATLAB/Simulink. Results: The behaviors of the design parameters of the HMT tractor in the working and driving modes were investigated as follows; For the stepwise change of the drawbar load in the working mode, the tractor and engine speeds were maintained at the desired values by the engine torque and HSU stroke control. In the driving mode, the tractor followed the desired speed through the control of the engine torque and HSU stroke. In this case, the engine operated near the OOL (optimal operating line) for the minimum fuel consumption within the shift range of HMT. Conclusions: A simulator for the HMT tractor was developed. The simulations were conducted under two operation conditions. It was found that the tractor speed and the engine speed are maintained at the desired values through the control of the engine torque and the HSU stroke.

To realize the high-speed intrusion detection by accommodating many regex-based signatures and the growing network link capacities, we propose a Service TimE-Aware Load balancing algorithm, which is called STEAL. This work is motivated from the observation that utilization of the many-core Network Intrusion Detection System (NIDS) is influenced by unfair computational distribution among many-core NIDS nodes. To avoid unfair computational distribution among many-core NIDS nodes, STEAL is designed to dynamically distribute the large volume of traffic among many-core NIDS nodes based on the packet service time, which is represented by the deep packet time in many-core NIDS nodes. From experiments, we show that compared to the commonly used load balancing algorithm based on arrival rate, STEAL increases the number of received packets, i.e., decreases the number of dropped packets, in many-core NIDS. Specifically, by integrating an open source NIDS, i.e. Bro, with STEAL, we show that even under the attack-dominant traffic and many signatures, STEAL can rapidly improve the performance of many-core NIDS to realize the high-speed intrusion detection.

Background: The injection pain of propofol is a frequent and well-known adverse effect. This study was designed to determine the optimal effect-site concentration of remifentanil for minimizing injection pain during induction with propofol. Methods: A total intravenous anesthetic technique was used for patients undergoing general anesthesia and remifentanil was pretreated to reach a certain target concentration before propofol injection. Using Dixon’s up-anddown method, the degree of pain described by the patient was used to adjust the target concentration of remifentanil for the next patient. Ten success-failure curves (crossovers) were sought to find the effect-site concentration (EC) of remifentanil for minimizing injection pain of propofol. Results: The EC of remifentanil in 50% and 95% of adult female population (EC 50 and EC95) for minimizing injection pain of propofol were 3.09 ng/ml (95% confidence limits [CI] 2.92 -3.30 ng/ml) and 3.78 ng/ml (95% CI 3.45-3.95 ng/ml), respectively. Clinically significant hemodynamic compromise or respiratory complications were not found during remifentanil infusion. Conclusions: Maintaining 3.78 ng/ml EC of remifentanil during induction with propofol attenuate propofol injection pain without serious adverse events in female patients undergoing general anesthesia and this method may provide the patient’s comfort without preparing other drugs for pain relief. (Korean J Anesthesiol 2012; 63: 108-112)

After a user logs into web server using a public personal computer or a mobile device such as mobile phone, the leakage of his personal information may occur because he often forgets to log out from web server. To prevent the leakage of the personal information instantaneously, we propose a new login status control method based on the connectivity of web servers and a single mobile application. The proposed method visualizes and timely controls login status information of a user. Thus, we can effectively preserve the user's personal information from accidentally or illegally being affected by anonymous users.

Background Subarachnoid block is a widely used technique for cesarean section. To improve the quality of analgesia and prolong the duration of analgesia, addition of intrathecal opioids to local anesthetics has been encouraged. We compared the effects of sufentanil 2.5 µg and 5 µg, which were added to intrathecal hyperbaric bupivacaine. Methods We enrolled 105 full term parturients were randomly divided into 3 groups: Group 1 (control), Group 2 (sufentanil 2.5 µg), and Group 3 (sufentanil 5 µg). In every group, 0.5% heavy bupivacaine was added according to the adjusted dose regimen. We determined the maximum level of sensory block and motor block, the quality of intraoperative analgesia, the duration of effective analgesia and side effects. Results There were no significant differences among the 3 groups in the maximum level of the sensory block and motor block. Recovery rate of the sensory block, however, was significantly slower in Group 3 than Group 1. Quality of intraopertive analgesia, muscle relaxation, and duration of effective analgesia were enhanced by increasing the dosage of intrathecal sufentanil. Frequencies of hypotension, maximum sedation level, and pruritus were directly related to the dosage of intrathecal sufentanil, whereas nausea and vomiting occurred only in the groups using sufentanil. Conclusions The addition of sufentanil 2.5 µg for spinal anesthesia provides adequate intraoperative analgesia and good postoperative analgesia with minimal adverse effects on the mother.

The flow between two concentric cylinders with the inner one rotating and with an imposed radial temperature gradient is studied using digital particle image velocimetry (DPIV) method. Four models of the outer cylinder without and with different numbers of slits (6, 9 and 18) are considered, and the radius ratio and aspect ratio of each models were 0.825 and 48, respectively. The flow regime in the Taylor-Couette flow was studied by increasing the Reynolds number. The results showed that smaller number of slits has no obvious effect on the transition process, which only change the shape of the vortex, and the transition to turbulent Taylor vortex is accelerated as the number of slit increases in both isothermal and non-isothermal conditions. It is also shown that the presence of temperature gradient increased the flow instability obviously as the Froude number larger than 0.0045.Copyright © 2009 by ASME

Background We report a case of a spray painter who developed malignant fibrous histiocytoma (MFH) of the maxillary sinus following long-term exposure to chromium, nickel, and formaldehyde, implying that these agents are probable causal agents of MFH. Case report The patient developed right-sided prosopalgia that began twenty months ago. The symptom persisted despite medical treatment. After two months, he was diagnosed with MFH through imaging studies, surgery, and pathological microscopic findings at a university hospital in Seoul. His social, medical, and family history was unremarkable. The patient had worked for about 18 years at an automobile repair shop as a spray painter. During this period, he had been exposed to various occupational agents, such as hexavalent chromium, nickel, and formaldehyde, without appropriate personal protective equipment. He painted 6 days a week and worked for about 8 hours a day. Investigation of the patient’s work environment detected hexavalent chromium, chromate, nickel, and formaldehyde. Conclusions The study revealed that the patient had been exposed to hexavalent chromium, formaldehyde, and nickel compounds through sanding and spray painting. The association between paranasal cancer and exposure to the aforementioned occupational human carcinogens has been established. We suggest, in this case, the possibility that the paint spraying acted as a causal agent for paranasal cancer.