• International Journal of Computer Science & Network Security
• /
• 제22권6호
• /
• pp.364-373
• /
• 2022

Image morphing methods make seamless transition changes in the image and mask the meaningful information attached to it. This can be detected by traditional machine learning algorithms and new emerging deep learning algorithms. In this research work, scope of different Hybrid learning approaches having combination of Deep learning and Machine learning are being analyzed with the public dataset CASIA V1.0, CASIA V2.0 and DVMM to find the most efficient algorithm. The simulated results with CNN (Convolution Neural Network), Hybrid approach of CNN along with SVM (Support Vector Machine) and Hybrid approach of CNN along with Random Forest algorithm produced 96.92 %, 95.98 and 99.18 % accuracy respectively with the CASIA V2.0 dataset having 9555 images. The accuracy pattern of applied algorithms changes with CASIA V1.0 data and DVMM data having 1721 and 1845 set of images presenting minimal accuracy with Hybrid approach of CNN and Random Forest algorithm. It is confirmed that the choice of best algorithm to find image forgery depends on input data type. This paper presents the combination of best suited algorithm to detect image morphing with different input datasets.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제50권7호
• /
• pp.339-349
• /
• 2001

This paper suggests an optimal identification method for complex and nonlinear system modeling that is based on Fuzzy-Neural Networks(FNN). The proposed Hybrid Identification Algorithm is based on Yamakawa's FNN and uses the simplified inference as fuzzy inference method and Error Back Propagation Algorithm as learning rule. In this paper, the FNN modeling implements parameter identification using HCM algorithm and hybrid structure combined with two types of optimization theories for nonlinear systems. We use a HCM(Hard C-Means) clustering algorithm to find initial apexes of membership function. The parameters such as apexes of membership functions, learning rates, and momentum coefficients are adjusted using hybrid algorithm. The proposed hybrid identification algorithm is carried out using both a genetic algorithm and the improved complex method. Also, an aggregated objective function(performance index) with weighting factor is introduced to achieve a sound balance between approximation and generalization abilities of the model. According to the selection and adjustment of a weighting factor of an aggregate objective function which depends on the number of data and a certain degree of nonlinearity(distribution of I/O data), we show that it is available and effective to design an optimal FNN model structure with mutual balance and dependency between approximation and generalization abilities. To evaluate the performance of the proposed model, we use the time series data for gas furnace, the data of sewage treatment process and traffic route choice process.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2016년도 춘계학술발표대회
• /
• pp.215-217
• /
• 2016

With the popularity of smart device, mobile applications are playing more and more important role in people's daily life, these applications stores various information which greatly facilitate the user's daily life. However due to the frequent transmission of data in the network also increases the risk of data leakage, more and more developers began to focus on how to protect user data. Current mainstream development models include Native development, Web development and Hybrid development. Hybrid development is based on JavaScript and HTML5, it has a cross platform advantages similar to Web Apps and a good user experience similar to Native Apps. In this paper according to the features of Hybrid applications, we proposed a security scheme in Hybrid development model implements message-level data encryption to protect user information. And through the performance evaluation we found that in some scenario the proposed security scheme has a better performance.

• 한국전산구조공학회논문집
• /
• 제23권2호
• /
• pp.123-137
• /
• 2010

하이브리드 구조실험에서는 전체구조물을 여러 개의 부분구조물로 나누어서 실험과 해석을 수행한다. 실험을 위한 부분 구조물들과 해석을 위한 부분구조물들은 지역적으로 서로 다른 장소에서 실험과 해석이 수행될 수 있으며, 이 부분구조물들의 실험과 해석은 시뮬레이션 코디네이터에 의하여 통제된다. 하이브리드 구조실험을 수행하는 동안에 시뮬레이션 코디네이터와 부분구조물들은 서로 간에 데이터 교환이 이루어지는 상호작용을 하게 된다. 본 논문은 이러한 상호작용을 기술하고 있는데, 하나의 하이브리드 구조실험 예제에 대하여 시뮬레이션 코디네이터와 부분구조물들 사이의 상호작용을 데이터 모델의 하나인 리하이 모델의 클래스와 객체를 통하여 표현하였다. 시뮬레이션 코디네이터와 부분구조물들은 각자의 데이터 저장을 위한 객체를 가지도록 구성하였고, 서로간의 연결은 동일한 형식의 인터페이스 링크를 사용하여 처리하였다. 본 논문에서 설명한 객체들은 일관된 방법에 의하여 구현하였는데, 하이브리드 구조실험을 위한 컴퓨터 시스템의 개발에 사용할 수 있다.

• 한국정보처리학회논문지
• /
• 제6권8호
• /
• pp.2031-2041
• /
• 1999

본 논문에서는 비정형, 대용량의 비디오데이터의 특징기반 검색과 주석기반 검색을 통합하여 다양한 사용자의 의미검색을 지원하고, 유사성 질의를 지원하는 통합비디오정보시스템(Hybrid Video Information System : HVIS)을 제안한다. HVIS는 메타데이터 모델링을 위해 한편의 비디오를 비디오 다큐먼트, 시퀸스, 장면, 객체로 나누고 물리적인 비디오스트림을 위한 원시데이터계층(raw_data layer)과 주석기반 검색, 특징기반 검색, 유사성 검색을 지원하기 위한 메타데이터계층(meta_data layer)의 두 개의 계층을 가진 통합 계층지향 메타데이터모델(Two layered Hybrid Object-oriented Metadata Model : THOMM)과 이 모델을 기반으로 주석기반 질의, 특징기반 질의, 유사질의가 가능한 비디오질의언어 (Video Query Language)와 질의를 처리하기 위한 비디오질의처리기 (Video Query Processor : VQP)와 질의처리알고리즘을 제안한다. 특히 유사한 장면, 객체를 찾는 유사질의시 사용자의 관심을 고려한 유사성 정도를 나타내는 식을 제시한다. 제안된 시스템은 Visual C++, ActiveX와 ORACLE를 이용하여 구현되었다.

• International Journal of Advanced Culture Technology
• /
• 제11권4호
• /
• pp.464-469
• /
• 2023

As modern society develops, the number of diseases caused by bad smells is increasing. As it can harm people's health, it is important to predict in advance the extent to which bad smells may occur, inform the public about this, and take preventive measures. In this paper, we propose a hybrid neural network structure of CNN and LSTM that can be used to detect or predict the occurrence of odors, which are most required in manufacturing or real life, using odor complex sensors. In addition, the proposed learning model uses a complex odor sensor to receive four types of data, including hydrogen sulfide, ammonia, benzene, and toluene, in real time, and applies this data to the inference model to detect and predict the odor state. The proposed model evaluated the prediction accuracy of the training model through performance indicators based on accuracy, and the evaluation results showed an average performance of more than 94%.

• 한국물환경학회지
• /
• 제31권5호
• /
• pp.507-522
• /
• 2015

In this study, some of the model verification results of STREAM (Spatio-Temporal River-basin Ecohydrology Analysis Model), a newly-developed hybrid watershed model, are presented for the runoff processes of nonpoint source pollution. For verification study of STREAM, the model was applied to a test watershed and a sensitivity analysis was also carried out for selected parameters. STREAM was applied to the Mankyung River Watershed to review the applicability of the model in the course of model calibration and validation against the stream flow discharge, suspended sediment discharge and some water quality items (TOC, TN, TP) measured at the watershed outlet. The model setup, simulation and data I/O modules worked as designed and both of the calibration and validation results showed good agreement between the simulated and the measured data sets: NSE over 0.7 and $R^2$ greater than 0.8. The simulation results also include the spatial distribution of runoff processes and watershed mass balance at the watershed scale. Additionally, the irrigation process of the model was examined in detail at reservoirs and paddy fields.

• International Journal of Concrete Structures and Materials
• /
• 제8권1호
• /
• pp.27-41
• /
• 2014

The use of carbon fibers (CF) and glass fibers (GF) were combined to strengthen concrete flexural members. In this study, data of tensile tests of 94 hybrid carbon-glass FRP sheets and 47 carbon and GF rovings or sheets were thoroughly investigated in terms of tensile behavior. Based on comparisons between the rule of mixtures and test data, positive hybrid effects were identified for various (GF/CF) ratios. Unlike the rule of mixtures, the hybrid sheets with relatively low (GF/CF) ratios also produced pseudo-ductility. From the calibrated results obtained from experiments, a new analytical model for the stress-strain relationship of hybrid FRP sheets was proposed. Finally, the hybrid effects were verified by structural tests of concrete members strengthened with hybrid FRP sheets and either carbon or glass FRP sheets.

• Journal of the Korean Data and Information Science Society
• /
• 제17권2호
• /
• pp.269-278
• /
• 2006

This paper presents the new hybrid data mining technique using error pattern, modeling of improving classification accuracy. The proposed method improves classification accuracy by combining two different supervised learning methods. The main algorithm generates error pattern modeling between the two supervised learning methods(ex: Neural Networks, Decision Tree, Logistic Regression and so on.) The Proposed modeling method has been applied to the simulation of 10,000 data sets generated by Normal and exponential random distribution. The simulation results show that the performance of proposed method is superior to the existing methods like Logistic regression and Discriminant analysis.

• Journal of the Korean Data and Information Science Society
• /
• 제26권1호
• /
• pp.271-280
• /
• 2015

In this paper, the maximum likelihood estimators for parameters are derived under three step-stress accelerated life tests for Type-I hybrid censored data. The exponential distribution and the cumulative exposure model are considered based on the assumption that a log quadratic relationship exits between stress and the mean lifetime ${\theta}$. The test plan to search optimal stress change times minimizing the asymptotic variance of maximum likelihood estimators are presented. A numerical example to illustrate the proposed inferential procedures and some simulation results to investigate the sensitivity of the optimal stress change times by the guessed parameters are given.