• 정보보호학회논문지
• /
• 제32권2호
• /
• pp.201-211
• /
• 2022

최근 사이버보안 패러다임의 변화에 따라, 인공지능 구현 기술인 기계학습과 딥러닝 기법을 적용한 이상탐지 방법의 연구가 증가하고 있다. 본 연구에서는 공개 데이터셋인 NGIDS-DS(Next Generation IDS Dataset)를 이용하여 GRU(Gated Recurrent Unit) 신경망 기반 침입 탐지 모델의 이상(anomaly) 탐지 성능을 향상시킬 수 있는 데이터 전처리 기술에 관한 비교 연구를 수행하였다. 또한 정상 데이터와 공격 데이터 비율에 따른 클래스 불균형 문제를 해결하기 위해 DCGAN(Deep Convolutional Generative Adversarial Networks)을 적용한 오버샘플링 기법 등을 사용하여 오버샘플링 비율에 따른 탐지 성능을 비교 및 분석하였다. 실험 결과, 시스템 콜(system call) 특성과 프로세스 실행패스 특성에 Doc2Vec 알고리즘을 사용하여 전처리한 방법이 좋은 성능을 보였고, 오버샘플링별 성능의 경우 DCGAN을 사용하였을 때, 향상된 탐지 성능을 보였다.

• Archives of Plastic Surgery
• /
• 제49권2호
• /
• pp.174-183
• /
• 2022

Management of traumatic skull base fractures and associated complications pose a unique reconstructive challenge. The goals of skull base reconstruction include structural support for the brain and orbit, separation of the central nervous system from the aerodigestive tract, volume to decrease dead space, and restoration of the three-dimensional appearance of the face and cranium with bone and soft tissues. An open bicoronal approach is the most commonly used technique for craniofacial disassembly of the bifrontal region, with evacuation of intracranial hemorrhage and dural repair performed prior to reconstruction. Depending on the defect size and underlying patient and operative factors, reconstruction may involve bony reconstruction using autografts, allografts, or prosthetics in addition to soft tissue reconstruction using vascularized local or distant tissues. The vast majority of traumatic anterior cranial fossa (ACF) injuries resulting in smaller defects of the cranial base itself can be successfully reconstructed using local pedicled pericranial or galeal flaps. Compared with historical nonvascularized ACF reconstructive options, vascularized reconstruction using pericranial and/or galeal flaps has decreased the rate of cerebrospinal fluid (CSF) leak from 25 to 6.5%. We review the existing literature on this uncommon entity and present our case series of n = 6 patients undergoing traumatic reconstruction of the ACF at an urban Level 1 trauma center from 2016 to 2018. There were no postoperative CSF leaks, mucoceles, episodes of meningitis, or deaths during the study follow-up period. In conclusion, use of pericranial, galeal, and free flaps, as indicated, can provide reliable and durable reconstruction of a wide variety of injuries.

• The Journal of Advanced Prosthodontics
• /
• 제14권4호
• /
• pp.212-222
• /
• 2022

PURPOSE. To evaluate and compare the effect of impression type (conventional vs digital) and the number of implants on the time from the impressions to the generation of working casts of mandibular implant-supported fixed completearch frameworks, as well as on patient satisfaction. MATERIALS AND METHODS. 17 participants, 3 or 4 implants, received 2 types of digital impression methods (DI) and conventional (CI). In DI, two techniques were performed: scanning with the scan bodies (SC) and scanning with a device attached to the scan bodies (SD) (BR 10 2019 026265 6). In CI, the making of a solid index (SI) and open-tray impression (OT) were used. The outcomes were used to evaluate the time and the participant satisfaction with conventional and digital impressions. The time was evaluated through the timing of the time obtained in the workflow in the conventional and digital impression. The effect of the number of implants on time was also assessed. Satisfaction was assessed through a questionnaire based on seven. The Wilcoxon test used to identify the statistical difference between the groups in terms of time. The Mann-Whitney test was used to analyze the relationship between the time and the number of implants. Fisher's test was used to assess the patient satisfaction (P<.05). RESULTS. The time with DI was shorter than with CI (DI, $\tilde{x}=02:58$; CI, $\tilde{x}=31:48$) (P<.0001). The arches rehabilitated with 3 implants required shorter digital impression time (3: $\tilde{x}=05:36$; 4: $\tilde{x}=09:16$) (P<.0001). Regarding satisfaction, the DI was more comfortable and pain-free than the CI (P<.005). CONCLUSION. Digital impressions required shorter chair time and had higher patient acceptance than conventional impressions.

• International Journal of Computer Science & Network Security
• /
• 제23권2호
• /
• pp.199-202
• /
• 2023

Carriage return (CR) and line feed (LF), also known as CRLF injection is a type of vulnerability that allows a hacker to enter special characters into a web application, altering its operation or confusing the administrator. Log poisoning and HTTP response splitting are two prominent harmful uses of this technique. Additionally, CRLF injection can be used by an attacker to exploit other vulnerabilities, such as cross-site scripting (XSS). Email injection, also known as email header injection, is another way that can be used to modify the behavior of emails. The Open Web Application Security Project (OWASP) is an organization that studies vulnerabilities and ranks them based on their level of risk. According to OWASP, CRLF vulnerabilities are among the top 10 vulnerabilities and are a type of injection attack. Automated testing can help to quickly identify CRLF vulnerabilities, and is particularly useful for companies to test their applications before releasing them. However, CRLF vulnerabilities can also lead to the discovery of other high-risk vulnerabilities, and it fosters a better approach to mitigate CRLF vulnerabilities in the early stage and help secure applications against known vulnerabilities. Although there has been a significant amount of research on other types of injection attacks, such as Structure Query Language Injection (SQL Injection). There has been less research on CRLF vulnerabilities and how to detect them with automated testing. There is room for further research to be done on this subject matter in order to develop creative solutions to problems. It will also help to reduce false positive alerts by checking the header response of each request. Security automation is an important issue for companies trying to protect themselves against security threats. Automated alerts from security systems can provide a quicker and more accurate understanding of potential vulnerabilities and can help to reduce false positive alerts. Despite the extensive research on various types of vulnerabilities in web applications, CRLF vulnerabilities have only recently been included in the research. Utilizing automated testing as a recurring task can assist companies in receiving consistent updates about their systems and enhance their security.

• 대한조선학회논문집
• /
• 제60권2호
• /
• pp.120-134
• /
• 2023

In this study, hydrodynamic characteristics of dam break flow were investigated by a series of experiments. The experiments were performed in a 2-D rectangular flume with obtaining instantaneous images of dam break flow to capture the free surface elevation, and pressure distributions on vertical wall and bottom of the flume. The initial water depth of the dam break flow was changed into 3 different heights, and the gate opening speed was changed during the experiments to study the effect of the gate speed in the dam break flow. Generation of dam break phenomena could be classified into three stages, i.e., very initial, relatively stable, and wall impact stages. The wall impact stage could be separated into 4 generation phases of wall impinge, run-up, overturning, and touchdown phases based on the deformation of the free surface. The free surface elevation were investigated with various initial water depth and compared with the analytic solutions by Ritter (1892). The pressures acting on the vertical wall and bottom were provided for the whole period of dam break flow varying the initial water depth and gate open speed. The measurement results of the dam break flow was compared with the hydrodynamic characteristics of green water phenomena, and it showed that the dam break flow could overestimate the green water loading based on the estimation suggested by Buchner (2002).

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제12권8호
• /
• pp.333-340
• /
• 2023

기업 부도의 효율적인 예측은 금융기관의 적절한 대출 결정과 여신 부실률 감소 측면에서 중요한 부분이다. 많은 연구에서 인공지능 기술을 활용한 분류모델 연구를 진행하였다. 금융 산업 특성상 새로운 예측 모델의 성능이 우수하더라도 어떤 근거로 결과를 출력했는지 직관적인 설명이 수반되어야 한다. 최근 미국, EU, 한국 등 에서는 공통적으로 알고리즘의 설명요구권을 제시하고 있어 금융권 AI 활용에 투명성을 확보하여야 한다. 본 논문에서는 외부에 오픈된 기업부도 데이터를 활용하여 인공지능 기반의 해석 가능한 분류 예측 모델을 제안하였다. 먼저 데이터 전처리 작업, 5겹 교차검증 등을 수행하고 로지스틱 회귀, SVM, XGBoost, LightGBM 등 10가지 지도학습 분류모델 최적화를 통해 분류 성능을 비교하였다. 그 결과 LightGBM이 가장 우수한 모델로 확인되었고, 설명 가능한 인공지능 기법인 SHAP을 적용하여 부도예측 과정에 대한 사후 설명을 제공하였다.

• 한국전자통신학회논문지
• /
• 제17권2호
• /
• pp.291-298
• /
• 2022

생성적 적대 신경망(Generative Adversarial Networks, GAN)은 내부의 두 신경망(생성망, 판별망)이 상호 경쟁하면서 학습하는 네트워크이다. 생성자는 현실과 가까운 이미지를 만들고, 구분자는 생성자의 이미지를 더 잘 감별하도록 프로그래밍 되어있다. 이 기술은 전체 이미지 X를 다른 이미지 Y로 생성, 변환 및 복원하기 위해 다양하게 활용되고 있다. 본 논문에서는 원본 이미지에서 부분 이미지만 추출한 후, 이를 자연스럽게 다른 객체로 위변조할 수 있는 방법에 관해 기술한다. 먼저 원본 이미지에서 부분 이미지만 추출한 후, 기존에 학습시켜놓은 DCGAN 모델을 통해 새로운 이미지를 생성하고, 이를 전체적 스타일 전이(overall style transfer) 기술을 사용하여 원본 이미지의 질감과 크기에 어울리도록 리스타일링(re-styling) 한 후, 원본 이미지에 자연스럽게 결합하는 과정을 거친다. 본 연구를 통해 원본 이미지의 특정 부분에 사용자가 원하는 객체 이미지를 자연스럽게 추가/변형할 수 있음으로써 가짜 이미지 생성의 또 다른 활용 분야로 사용될 수 있을 것이다.

• International Journal of Advanced Culture Technology
• /
• 제10권1호
• /
• pp.274-283
• /
• 2022

Korea is facing a number difficulties arising from rising housing prices. As 'housing' takes the lion's share in personal assets, many difficulties are expected to arise from fluctuating housing prices. The purpose of this study is creating housing price prediction model to prevent such risks and induce reasonable real estate purchases. This study made many attempts for understanding real estate instability and creating appropriate housing price prediction model. This study predicted and validated housing prices by using the LSTM technique - a type of Artificial Intelligence deep learning technology. LSTM is a network in which cell state and hidden state are recursively calculated in a structure which added cell state, which is conveyor belt role, to the existing RNN's hidden state. The real sale prices of apartments in autonomous districts ranging from January 2006 to December 2019 were collected through the Ministry of Land, Infrastructure, and Transport's real sale price open system and basic apartment and commercial district information were collected through the Public Data Portal and the Seoul Metropolitan City Data. The collected real sale price data were scaled based on monthly average sale price and a total of 168 data were organized by preprocessing respective data based on address. In order to predict prices, the LSTM implementation process was conducted by setting training period as 29 months (April 2015 to August 2017), validation period as 13 months (September 2017 to September 2018), and test period as 13 months (December 2018 to December 2019) according to time series data set. As a result of this study for predicting 'prices', there have been the following results. Firstly, this study obtained 76 percent of prediction similarity. We tried to design a prediction model of real estate transaction price with the LSTM Model based on AI and Bigdata. The final prediction model was created by collecting time series data, which identified the fact that 76 percent model can be made. This validated that predicting rate of return through the LSTM method can gain reliability.

• Structural Engineering and Mechanics
• /
• 제85권4호
• /
• pp.433-443
• /
• 2023

The current paper discusses the dynamic and stability responses of cross-ply composite laminated plates by employing a refined quasi-3D trigonometric shear deformation theory. The proposed theory takes into consideration shear deformation and thickness stretching by a trigonometric variation of in-plane and transverse displacements through the plate thickness and assures the vanished shear stresses conditions on the upper and lower surfaces of the plate. The strong point of the new formulation is that the displacements field contains only 4 unknowns, which is less than the other shear deformation theories. In addition, the present model considers the thickness extension effects (ε_z≠0). The presence of the Winkler-Pasternak elastic base is included in the mathematical formulation. The Hamilton's principle is utilized in order to derive the four differentials' equations of motion, which are solved via Navier's technique of simply supported structures. The accuracy of the present 3-D theory is demonstrated by comparing fundamental frequencies and critical buckling loads numerical results with those provided using other models available in the open literature.

• Restorative Dentistry and Endodontics
• /
• 제47권4호
• /
• pp.36.1-36.17
• /
• 2022

Objectives: This study aimed to evaluate the effect of aging on the marginal quality of glass hybrid (GH) material used to elevate dentin gingival margins, and to analyze the consistency of the results obtained by 3 in vitro methods. Materials and Methods: Ten teeth received compound class II cavities with subgingival margins. The dentin gingival margins were elevated with GH, followed by resin composite. The GH/gingival dentin interfaces were examined through digital microscopy, scanning electron microscopy (SEM) using resin replicas, and according to the World Dental Federation (FDI) criteria. After initial evaluations, all teeth were subjected to 10,000 thermal cycles, followed by repeating the same marginal evaluations and energy dispersive spectroscopy (EDS) analysis for the interfacial zone of 2 specimens. Marginal quality was expressed as the percentage of continuous margin at ×200 for microscopic techniques and as the frequency of each score for FDI ranking. Data were analyzed using the paired sample t-test, Wilcoxon signed-rank test, and Pearson and Spearmen correlation coefficients. Results: None of the testing techniques proved the significance of the aging factor. Moderate and strong significant correlations were found between the testing techniques. The EDS results suggested the presence of an ion-exchange layer along the GH/gingival dentin interface of aged specimens. Conclusions: The marginal quality of the GH/dentin gingival interface defied aging by thermocycling. The replica SEM and FDI ranking results had stronger correlations with each other than either showed with the digital microscopy results.