• The Transactions of the Korea Information Processing Society
• /
• v.2 no.6
• /
• pp.847-856
• /
• 1995

When a human recognizes nouns in a sentence, s/he associates them with the hyper concepts of onus. For computer to simulate the human's word recognition, it should build the knowledge base (WordNet)for the hyper concepts of words. Until now, works for the WordNet haven't been performed in Korea, because they need lots of human efforts and time. But, as the power of computer is radically improved and common MRD becomes available, it is more feasible to automatically construct the WordNet. This paper proposes the method that automatically builds the WordNet of Korean nouns by using the descripti on of onus in Korean MRD, and it proposes the rules for extracting the hyper concepts (hypernyms)by analyzing structrual characteristics of Korean. The rules effect such characteristics as a headword lies on the rear part of sentences and the descriptive sentences of nouns have special structure. In addition, the WordNet prototype of Korean Nouns is developed, which is made by combining the hypernyms produced by the rules mentioned above. It extracts the hypernyms of about 2,500 sample words, and the result shows that about 92per cents of hypernyms are correct.

• ETRI Journal
• /
• v.39 no.4
• /
• pp.502-513
• /
• 2017

For video understanding, namely analyzing who did what in a video, actions along with objects are primary elements. Most studies on actions have handled recognition problems for a well-trimmed video and focused on enhancing their classification performance. However, action detection, including localization as well as recognition, is required because, in general, actions intersect in time and space. In addition, most studies have not considered extensibility for a newly added action that has been previously trained. Therefore, proposed in this paper is an extensible hierarchical method for detecting generic actions, which combine object movements and spatial relations between two objects, and inherited actions, which are determined by the related objects through an ontology and rule based methodology. The hierarchical design of the method enables it to detect any interactive actions based on the spatial relations between two objects. The method using object information achieves an F-measure of 90.27%. Moreover, this paper describes the extensibility of the method for a new action contained in a video from a video domain that is different from the dataset used.

• Journal of Korean Association for Spatial Structures
• /
• v.20 no.2
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• The KIPS Transactions:PartB
• /
• v.13B no.4 s.107
• /
• pp.439-448
• /
• 2006

Previous recognition/clustering algorithms such as Kohonen SOM(Self-Organizing Map), MLP(Multi-Layer Percecptron) and SVM(Support Vector Machine) might not adapt to unexpected input pattern. And it's recognition rate depends highly on the complexity of own training patterns. We could make up for and improve the weak points with lowering complexity of original problem without losing original characteristics. There are so many ways to lower complexity of the problem, and we chose a kernel concepts as an approach to do it. In this paper, using a kernel concepts, original data are mapped to hyper-dimension space which is near infinite dimension. Therefore, transferred data into the hyper-dimension are distributed spasely rather than originally distributed so as to guarantee the rate to be risen. Estimating ratio of recognition is based on a new similarity-probing and learning method that are proposed in this paper. Using CEDAR DB which data is written in cursive letters, 0 to 9, we compare a recognition/clustering performance of kSOM that is proposed in this paper with previous SOM.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.34 no.2
• /
• pp.175-181
• /
• 2010

This paper presents the dynamic analysis method for estimating the performance of flat-type blades in wiper systems. The blade has nonlinear characteristics since the rubber is a hyper-elastic material. Thus, modal coordinate and absolute nodal coordinate formulations were used to describe the dynamic characteristic of the blade. The blade was structurally analyzed to find the bending characteristics of the cross section of the blade. According to the analysis results, the blade section is divided into three deformation bodies: rigid, small, and large. For the small deformation body, the modal coordinate formulation is used, while the absolute nodal coordinate formulation is used for the large deformation body. To verify the dynamic analysis result, an experiment was performed. The simulation and experiment results were compared to verify the flexible multi-body dynamic model.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.3
• /
• pp.985-989
• /
• 2012

Objective: To make sure the feasibility with $^{18F}FDG$ PET/CT to guided dynamic intensity-modulated radiation therapy (IMRT) for nasopharyngeal carcinoma patients, by dosimetric verification before treatment. Methods: Chose 11 patients in III~IVA nasopharyngeal carcinoma treated with functional image-guided IMRT and absolute and relative dosimetric verification by Varian 23EX LA, ionization chamber, 2DICA of I'mRT Matrixx and IBA detachable phantom. Drawing outline and making treatment plan were by different imaging techniques (CT and $^{18F}FDG$ PET/CT). The dose distributions of the various regional were realized by SMART. Results: The absolute mean errors of interest area were $2.39%{\pm}0.66$ using 0.6cc ice chamber. Results using DTA method, the average relative dose measurements within our protocol (3%, 3 mm) were 87.64% at 300 MU/min in all filed. Conclusions: Dosimetric verification before IMRT is obligatory and necessary. Ionization chamber and 2DICA of I'mRT Matrixx was the effective dosimetric verification tool for primary focal hyper metabolism in functional image-guided dynamic IMRT for nasopharyngeal carcinoma. Our preliminary evidence indicates that functional image-guided dynamic IMRT is feasible.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.9
• /
• pp.1915-1922
• /
• 2012

In this paper, the Solar Inverter will be monitored by using the AJAX(Asynchronous JavaScript and XML). AJAX is the one of the technologies that can make the RIA(Rich Internet Application) with DHTML(Dynamic Hyper Text Makeup Language) and other java script technology. By using this, a strong application program that is comparable to the general application program can be made. With an existing data-processing technique, the request and response of data can't be processed dynamically on the same page. However, real-time monitoring of data and operation statuses can be confirmed by using the AJAX an asynchronous method of communication. Also without changing the page, the amount of data transmission used the AJAX with significantly small amounts of data to build a Solar Inverter monitoring system that is able to efficiently handle management and monitoring, operating all functions within one page.

• The KIPS Transactions:PartC
• /
• v.9C no.6
• /
• pp.865-874
• /
• 2002

HTTP (HyperText Transfer Protocol) is a transfer protocol used by the World Wide Web distributed hypermedia system to retrieve the objects. Because HTTP is a connection oriented protocol, it uses TCP (Transmission control Protocol) as a transport layer. But it is known that HTTP interacts with TCP badly. it is discussed about factors affecting the performance or HTTP over TCP, the transaction time obtained by the per-transaction TCP connections for HTTP access and the TCP slow-start overheads, and the transaction time for T-TCP (Transaction TCP) which is one or methods improving the performance or HTTP over TCP. Average waiting time and average turnaround time are important parameters to satisfy QoS (Quality of Service) of end users. Formulas for calculating two parameters are derived. Such formulas can be used for the environment in which each TCP or T-TCP transaction time is same or different. Some experiments and computational experiences indicate that the proposed formulas are well acted, can be applied to the environment which the extension of bandwidth is necessary, and time characteristics of T-TCP are superior to that of TCP. Also, the load distribution method of web server based on the combination of bandwidths is discussed to reduce average waiting time and average turnaround time.

• Coupled systems mechanics
• /
• v.4 no.4
• /
• pp.279-295
• /
• 2015

The article presents a theoretical-experimental approach developed for modeling the coefficient of sliding friction in the dynamic system tool-workpiece in slide diamond burnishing of low-alloy unhardened steels. The experimental setup, implemented on conventional lathe, includes a specially designed device, with a straight cantilever beam as body. The beam is simultaneously loaded by bending (from transverse slide friction force) and compression (from longitudinal burnishing force), which is a reason for geometrical nonlinearity. A method, based on the idea of separation of the variables (time and metric) before establishing the differential equation of motion, has been applied for dynamic modeling of the beam elastic curve. Between the longitudinal (burnishing force) and transverse (slide friction force) forces exists a correlation defined by Coulomb's law of sliding friction. On this basis, an analytical relationship between the beam deflection and the sought friction coefficient has been obtained. In order to measure the deflection of the beam, strain gauges connected in a "full bridge" type of circuit are used. A flexible adhesive is selected, which provides an opportunity for dynamic measurements through the constructed measuring system. The signal is proportional to the beam deflection and is fed to the analog input of USB DAQ board, from where the signal enters in a purposely created virtual instrument which is developed by means of Labview. The basic characteristic of the virtual instrument is the ability to record and visualize in a real time the measured deflection. The signal sampling frequency is chosen in accordance with Nyquist-Shannon sampling theorem. In order to obtain a regression model of the friction coefficient with the participation of the diamond burnishing process parameters, an experimental design with 55 experimental points is synthesized. A regression analysis and analysis of variance have been carried out. The influence of the factors on the friction coefficient is established using sections of the hyper-surface of the friction coefficient model with the hyper-planes.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.12
• /
• pp.14-24
• /
• 2018

Several typical hyper-elastic constitutive models that encompass both conventional and advanced ones were investigated for the application of instability problems, including the biaxial tension of a rubber patch and inflation of spherical or cylindrical balloons. The material models included the neo-Hookean model, Mooney-Rivlin model, Gent model, Arruda-Boyce model, Fung model, and Pucci-Saccomandi model. Analyses can be done using membrane equations with particular strain energy density functions. Among the typical strain energy density functions, Kearsley's bifurcation for the Treloar's patch occurs only with the Mooney-Rivlin model. The inflation equation is so generalized that a spherical balloon and tube balloons can be taken into account. From the analyses, the critical material parameters and limit points were identified for material models in terms of the non-dimensional pressure and inflation volume ratio. The bifurcation was then identified and found for each material model of a balloon. When the finite element method was used for the structural instability problems of rubber-like materials, some careful treatments required could be suggested. Overall, care must be taken not only with the analysis technique, but also in selecting constitutive models, particularly the instabilities.