Search

In this study, we focus on real-time anomaly detection using the gated graph neural network (GGNN) and long short-term memory (LSTM) algorithms for the most commonly used network protocols of HTTP and TELNET. In a network, the flow of the various protocols is determined by their respective roles. Each protocol consists of options with individual tasks, and protocols are processed based on these tasks. Therefore, when analyzing flow, a unique repetitive pattern emerges according to the flow and mission. Accordingly, when an anomaly signal is involved, the pattern shows different characteristics from the existing pattern. This study identifies the flow of input and output values based on the correlation between the options defined at each port, whereby the correlation is analyzed, and the detection accuracy of the anomaly signal is determined using the GGNN and LSTM algorithms. The experimental results demonstrate an accuracy of 99.36% for GGNN and 89.46% for LSTM in detecting network anomalies. [ABSTRACT FROM AUTHOR]

Various studies have been conducted to detect network anomalies. However, because anomaly signals are determined by the pattern characteristics using the dataset, the real-time detection problem continues. Even if there is a signal with an attack sign among the constantly transmitted and received signals, the attack cannot be blocked in advance. Moreover, it appears in many places in a distributed denial-of-service (DDoS) attack, so the real-time defense must be the best option. Therefore, it is necessary first to discover the characteristics and elements regarded as abnormal signals to discover anomalies in real time. Finally, by analyzing the correlation between network data and features, extracting the elements of the anomaly, and analyzing the behavior of the extracted elements in detail, we aim to increase the accuracy of the anomaly. In this study, we used Coburg intrusion detection and KDDCup datasets and analyzed the correlation of elements in the dataset using a graph neural network. The calculated accuracy values of the anomaly detection were 94.5% and 98.85%. [ABSTRACT FROM AUTHOR]

The Baekje Muryeong royal tomb is located in Gongju City, Republic of Korea. It is unearthed in its entirety and can confirm accurate information. Various artifacts such as grave name and Jinmyosoo have been discovered and provide important data for Baekje research. It is desirable to provide experience through virtual reality because access is limited for preservation and restoration. The construction, completion, and excavation of Muryeong Royal tomb are composed of virtual reality contents. The elements of virtual reality consist of navigation, interaction, immersion, reality, content, and community. Experiential contents and VR components were presented in correspondence. These results are a fundamental task for content development that can focus on the inherent value of the heritage. [ABSTRACT FROM AUTHOR]

Photoplethysmography imaging (PPGI) sensors have attracted a significant amount of attention as they enable the remote monitoring of heart rates (HRs) and thus do not require any additional devices to be worn on fingers or wrists. In this study, we mounted PPGI sensors on a robot for active and autonomous HR (R-AAH) estimation. We proposed an algorithm that provides accurate HR estimation, which can be performed in real time using vision and robot manipulation algorithms. By simplifying the extraction of facial skin images using saturation (S) values in the HSV color space, and selecting pixels based on the most frequent S value within the face image, we achieved a reliable HR assessment. The results of the proposed algorithm using the R-AAH method were evaluated by rigorous comparison with the results of existing algorithms on the UBFC-RPPG dataset (n = 42). The proposed algorithm yielded an average absolute error (AAE) of 0.71 beats per minute (bpm). The developed algorithm is simple, with a processing time of less than 1 s (275 ms for an 8-s window). The algorithm was further validated on our own dataset (BAMI-RPPG dataset [n = 14]) with an AAE of 0.82 bpm. [ABSTRACT FROM AUTHOR]

The small cell structure in which many cells are arranged per unit area by reducing the size of cells is a candidate technology for an increase in transmission capacity in the 5G environment. However, the decrease in the size of the cell led to additional problems such as increased inter cell interference and frequent cell changes owing to the movement of the terminal. Therefore, the aim of this study was to propose small cell dynamic channel allocation (SDCA) and hybrid and dynamic channel allocation (HDCA) using conventional reuse methods to improve the macro cell performance while efficiently utilizing scarce frequency resources. The proposed method facilitates an improved performance that is lacking for macro-cell users in the center area of the cell boundary for the network where conventional macro cells and small cells are superposed. Furthermore, to improve the performance, it can provide resources that are lacking in the small cells of the center. To evaluate the performance, the proposed method was compared to frequency reuse factor1 (FRF1), frequency reuse factor3 (FRF3), and fractional frequency reuse (FFR) methods in terms of the signal-to-interference/noise-ratio (SINR) of users of each macro cell and small cell, outage, capacity for each user, and total system capacity. As a result of comparing the SINR, it was confirmed that the performance of the macro cell users has improved by an average of 43.88% compared to FRF1, FRF3, and FFR, and the performance of small cell users has improved by an average of 4.31%. Comparison results show that the outage proportions of the macro and small cell users are 61.29% and 70.59% lower on average, respectively. A comparison of results show that the capacities of the macro and small cell users have also improved by 22.5% and 14.5% on average, respectively. As the comparison results of the total system capacity indicate, the proposed method shows an average improvement of 11.67%. In cases in which the added resources of the small cells are found to be unnecessary based on the results of the performance evaluation, there is an advantage in that they can be reduced to improve the performance of macro cell users, or they can be used to fill the insufficient resources of the small cells while maintaining the performance of the macro cell users. This fluidity originates from the ability to address occasional situations in a dense environment. These two approaches are expected to be used effectively in 5G network environments. [ABSTRACT FROM AUTHOR]

The aim of the study is to develop artificial intelligence (AI) algorithm based on a deep learning model to predict mortality using abbreviate injury score (AIS). The performance of the conventional anatomic injury severity score (ISS) system in predicting in-hospital mortality is still limited. AIS data of 42,933 patients registered in the Korean trauma data bank from four Korean regional trauma centers were enrolled. After excluding patients who were younger than 19 years old and those who died within six hours from arrival, we included 37,762 patients, of which 36,493 (96.6%) survived and 1269 (3.4%) deceased. To enhance the AI model performance, we reduced the AIS codes to 46 input values by organizing them according to the organ location (Region-46). The total AIS and six categories of the anatomic region in the ISS system (Region-6) were used to compare the input features. The AI models were compared with the conventional ISS and new ISS (NISS) systems. We evaluated the performance pertaining to the 12 combinations of the features and models. The highest accuracy (85.05%) corresponded to Region-46 with DNN, followed by that of Region-6 with DNN (83.62%), AIS with DNN (81.27%), ISS-16 (80.50%), NISS-16 (79.18%), NISS-25 (77.09%), and ISS-25 (70.82%). The highest AUROC (0.9084) corresponded to Region-46 with DNN, followed by that of Region-6 with DNN (0.9013), AIS with DNN (0.8819), ISS (0.8709), and NISS (0.8681). The proposed deep learning scheme with feature combination exhibited high accuracy metrics such as the balanced accuracy and AUROC than the conventional ISS and NISS systems. We expect that our trial would be a cornerstone of more complex combination model. [ABSTRACT FROM AUTHOR]

The risk of tampering exists for conventional user recognition methods based on biometrics such as face and fingerprint. Recently, research on user recognition using biometric signals such as electrocardiogram (ECG), electroencephalogram (EEG), and electromyogram (EMG) has been actively performed to overcome this issue. We herein propose a user recognition method applying a deep learning technique based on ensemble networks after transforming ECG signals into two-dimensional (2D) images. A preprocessing process for one-dimensional ECG signals is performed to remove noise or distortion; subsequently, they are projected onto a 2D image space and transformed into image data. For the proposed algorithm, we designed deep learning-based ensemble networks to improve the degraded performance arising from overfitting in a single network. Our experimental results demonstrate that the proposed ensemble networks exhibit an accuracy that is 1.7% higher than that of the single network. In particular, the performance of the ensemble networks is up to 13% higher compared to the single network that degrades the recognition rate by displaying similar features between classes. [ABSTRACT FROM AUTHOR]

Prolonged pleural effusion after Fontan operation is a significant morbidity that leads to long hospital stays. We investigated the association of multiple risk factors, including clinical characteristics, hemodynamic parameters, and preoperative, operative, and postoperative factors, with prolonged pleural effusion after Fontan operation. Eighty-five patients who underwent a Fontan operation between January 2005 and June 2018 in our center were included in this retrospective study. Patients were divided into two groups: group 1 (n = 36, 42.4%) included those with prolonged pleural effusion, defined as lasting > 14 days after the Fontan operation, and group 2 included patients without prolonged pleural effusion. Patients with hypoplastic left heart syndrome (HLHS) were more prevalent in group 1 (n = 15, P = 0.006). No differences in age at Fontan operation, central venous pressure at Fontan operation, or hemodynamic parameters during the pre-Fontan evaluation were found between the two groups. In multivariable analysis, HLHS (P = 0.002), non-fenestration (P = 0.018), and high central venous pressure at bidirectional cavopulmonary shunt (BCPS) operation (P = 0.043) were independent risk factors for prolonged pleural effusion after Fontan operation. Adverse outcomes such as death, need for heart transplantation, and Fontan failure were not associated with prolonged pleural effusion. In conclusion, patients with HLHS and higher central venous pressure at BCPS were more likely to have a prolonged pleural effusion after Fontan operation, but fenestration was more likely to decrease prolonged effusion. We should consider closer management of fluid status before, during, and after surgery in patients with these risk factors after Fontan operation. [ABSTRACT FROM AUTHOR]

Coronary reimplantation after neoaortic reconstruction (CRANR) in the arterial switch operation (ASO) allows easy selection of accurate coronary transfer sites in the distended neoaorta. However, neoaortic valve injury may occur during coronary reimplantation. We determined whether the CRANR procedure increased the incidence of aortic valve regurgitation (AR) after ASO. Between March 1994 and August 2017, 227 patients underwent ASO. Since September 2000 CRANR has been performed on 155 patients and open coronary reimplantation (OCR) on 72. Patients who had undergone aortocoronary flaps procedures (n = 13), had early or late mortality (n = 27), or lacked data (n = 11) were excluded. We enrolled and retrospectively reviewed the medical records of 176 patients who were followed up for postoperative AR: 38 underwent OCR and 138 underwent CRANR. We compared the incidences of early and late postoperative AR in both groups. We defined mild or greater AR as "significant AR." The groups did not differ in body weight at operation, great artery relationship, and coronary artery anatomy. The incidences of significant AR at discharge were 21.1% (8/38) in the OCR group and 16.6% (23/138) in the CRANR group (p = 0.53). The freedom from significant AR at 5 years was 59.9% in the OCR group and 62.4% in the CRANR group with no difference between the two groups (p = 0.73). Moderate AR occurred in one patient in the CRANR group. No surgical intervention was required for the aortic valve in either group. ASO using the CRANR technique did not increase the incidence of postoperative early and late AR. [ABSTRACT FROM AUTHOR]

An introduction is presented in which the editor discusses articles in the issue on topics including smartphone sensory data; genetic algorithms; and pattern recognition.

A Bayesian network is a promising probabilistic model to represent causal relations between nodes (random variables). One of the major research issue in a Bayesian network is how to infer causal relations from a dataset by constructing better heuristic learning algorithms. Many kinds of approaches were so far introduced, and estimation of distribution algorithms (EDAs) are one of the promising causal discovery algorithms. However, the performance of EDAs is considerably dependent on the quality of the first population because new individuals are reproduced from the previous populations. In this paper, we introduce a new initialization method for EDAs that extracts promising candidate causal relations based on causal scores. Then, we used the promising relations to construct a better first population and to reproduce better individuals until the learning algorithm is terminated. Experimental results show that EDAs infer a more number of correct causal relations when promising relations were used in EDA based structure learning. It means that the performance of EDAs can be improved by providing better local search space, and it was the promising relations in this paper. [ABSTRACT FROM AUTHOR]

The management of pulmonary atresia with a ventricular septal defect (PA/VSD) depends on the anatomy of the pulmonary artery or on the surgical strategy used at individual institutions. In our institution, we adopted a right ventricle-to-pulmonary artery (RV-PA) shunt in 2011 as a palliative procedure for PA/VSD to overcome the disadvantages of a Blalock-Taussig shunt. We evaluated the results of the RV-PA shunt as initial palliative surgery for PA/VSD. Thirteen patients with ductus-dependent PA/VSD from August 2011 to August 2015 were enrolled. The mean age at surgery was 17.9 ± 15.3 (range 5-60) days, and the mean body weight was 2.9 ± 0.6 (range 2.2-4.0) kg. A Gore-Tex tube graft was used in all patients. We retrospectively observed intra- and postoperative complications, early and late mortality, and palliation duration to definitive repair. Left pulmonary artery angioplasty was performed as a concomitant procedure in three patients. There were no early hospital mortalities, although two inter-stage deaths occurred 34 and 47 days postoperatively: one patient died of aspiration and the other of right ventricular outflow tract (RVOT) pseudoaneurysm rupture. Two patients (15.4%) required the extracorporeal membrane oxygenation support postoperatively: one because of failure to wean from a bypass caused by persistent hypoxemia and the other because of sudden massive bleeding from the RVOT suture line in the intensive care unit 2 days postoperatively. These two patients underwent second-stage definitive repair successfully. Five patients (41.7%) required catheter intervention, for juxtaductal left pulmonary artery stenosis in three patients, right pulmonary artery stenosis in one, and shunt inflow stenosis in one. Two patients (15.4%) required re-operation because of shunt inflow stenosis and RVOT pseudoaneurysm, respectively. All patients who survived the RV-PA shunt underwent total correction at a mean interval of 13.1 months. A RV-PA shunt is an option for the initial palliation of ductus-dependent PA/VSD. Major complications can occur, including RVOT pseudoaneurysm, shunt inflow stenosis, persistent hypoxemia during the immediate postoperative period, and dehiscence of the shunt anastomosis site. Caution should be taken when performing the RV-PA shunt for palliation of PA/VSD. [ABSTRACT FROM AUTHOR]

To measure the similarity of contexts in smart devices, comparison is made of the user-defined contexts and another context which is defined in a server or in a network device that a user has (Segev and Toch IEEE Trans Serv Comput 2(3):210-222, 2009). While it processes to compare, if they find some users who have similarities with them, they surely may be interested in the users, because they know that they can share their information without the time wasting to search, finally getting what they want. However, according to the characteristics of the registered contexts, they are classified into two types, a rank-definitional context and a rank-undefined context. Also, the users usually want to use the two types to get quickly what they want at the same time. It means that another algorithm may be needed to get the similarity depending on the contexts, because the existing similarity search algorithms usually perform the similarity process without the contexts' characteristics analysis. They assume that all contexts have the same features when they process. Now, the existing methods that find the similarity usually have an accuracy problem. Low accuracy gives invisible services to users. This paper has suggested named context-based pattern measurement method including weight defines for higher accuracy. As a result, it would be able to get accuracy similarity by applying to the proposed algorithms about 69.072 % without weight and also 95.322 % accuracy in case it has a specific weight. [ABSTRACT FROM AUTHOR]

Background: Image artifacts affect the quality of medical images and may obscure anatomic structure and pathology. Numerous methods for suppression and correction of scattered image artifacts have been suggested in the past three decades. In this paper, we assessed the feasibility of use of information on scattered artifacts for estimation of bone mineral density (BMD) without dual-energy X-ray absorptiometry (DXA) or quantitative computed tomographic imaging (QCT). Methods: To investigate the relationship between scattered image artifacts and BMD, we first used a forearm phantom and cone-beam computed tomography. In the phantom, we considered two regions of interest-bone-equivalent solid material containing 50 mg HA per cm and water-to represent low- and high-density trabecular bone, respectively. We compared the scattered image artifacts in the high-density material with those in the low-density material. The technique was then applied to osteoporosis patients and healthy subjects to assess its feasibility for BMD estimation. Results: The high-density material produced a greater number of scattered image artifacts than the low-density material. Moreover, the radius and ulna of healthy subjects produced a greater number of scattered image artifacts than those from osteoporosis patients. Conclusions: Although other parameters, such as bone thickness and X-ray incidence, should be considered, our technique facilitated BMD estimation directly without DXA or QCT. We believe that BMD estimation based on assessment of scattered image artifacts may benefit the prevention, early treatment and management of osteoporosis. [ABSTRACT FROM AUTHOR]

Retraction Note to: Telecommunication Systems (2002) 80:189-214 https://doi.org/10.1007/s1123... The authors have retracted this article because it contains material that substantially overlaps with the following thesis [[1]]. A. K. Singh has not responded to any correspondence from the editor/publisher about this retraction. [Extracted from the article]

Despite of various benefits such as a convenience and efficiency, home healthcare systems have some inherent security risks that may cause a serious leak on personal health information. This work presents a Secure User Profiling Structure which has the patient information including their health information. A patient and a hospital keep it at that same time, they share the updated data. While they share the data and communicate, the data can be leaked. To solve the security problems, a secure communication channel with a hash function and an One-Time Password between a client and a hospital should be established and to generate an input value to an OTP, it uses a dual hash-function. This work presents a dual hash function-based approach to generate the One-Time Password ensuring a secure communication channel with the secured key. In result, attackers are unable to decrypt the leaked information because of the secured key; in addition, the proposed method outperforms the existing methods in terms of computation cost. [ABSTRACT FROM AUTHOR]