• Proceedings of the ESK Conference
• /
• 1996.04a
• /
• pp.36-47
• /
• 1996

By integrating the eye head-position monitioring devices, the present authors developed an eye-controlled human/computer interface based on the line-of-sight and an intentional blink to invoke commands. Also modified was an existing calibration method to reduce the visual angle between the target center and the intersection point of the derived line-of-sight. This modified calibration method allowed 108 or more command blocks to be displayed on the 14 inch monitor with the target acquisition probability(hit rate) of 98% when viewed at the distance of 500 mm apart. An active triggering method using an intentional blink was proposed and was shown to be a feasible and efficient alternative to invoke commands with total triggering time of 0.8 sec or less. The system could be used by the normal people as well as the handicapped individuals as a new human/computer interface.

• Journal of Mechanical Science and Technology
• /
• v.20 no.8
• /
• pp.1248-1255
• /
• 2006

The velocity and pressure fields of a ship's propulsion mechanism of the Weis-Fogh type, in which a airfoil moves reciprocally in a channel, are studied in this paper using the advanced vortex method. The airfoil and the channel are approximated by a finite number of source and vortex panels, and the free vortices are introduced from the body surfaces. The viscous diffusion of fluid is represented using the core-spreading model to the discrete vortices. The velocity is calculated on the basis of the generalized Biot-Savart law and the pressure field is calculated from integrating the equation given by the instantaneous velocity and vorticity fields. Two-dimensional unsteady viscose flows of this propulsion mechanism are numerically clarified, and the calculated results agree well with the experimental ones.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2004.05b
• /
• pp.370-373
• /
• 2004

We must have management method of schema conflict for integrating local schema on content integration and management system. But The past method is inconsistent and depends on local and global databases. In this paper, we propose the management method of schema confilct using XML that can represent and interchange the structual information of schema

• Proceedings of the KIEE Conference
• /
• 2005.07b
• /
• pp.1204-1206
• /
• 2005

This paper deals with magnetic field analysis and computation of cogging torque in IPM motor with an analytical method, which is based on the Conformal Mapping technique. The magnetic field is analyzed by solving space harmonic field analysis due to inserted PM magnetizing distribution. Conformal Mapping method is then used for considering the slot opening effect and rotor saliency effect on the air-gap field magnetic distribution. Then, by integrating the field over the stator surface, cogging torque is calculated. The validity of the proposed analytical method is confirmed by comparing the results with 2-D FEA results.

• Korean Journal of Remote Sensing
• /
• v.26 no.6
• /
• pp.693-703
• /
• 2010

High spectral resolution of hyperspectral data enables analysis of complex natural phenomena that is reflected on the data nonlinearly. Although many manifold learning methods have been developed for such problems, most methods do not consider the spatial correlation between samples that is inherent and useful in remote sensing data. We propose a manifold learning method which directly combines the spatial proximity and the spectral similarity through kernel PCA framework. A gain factor caused by spatial proximity is first modelled with a heat kernel, and is added to the original similarity computed from the spectral values of a pair of samples. Parameters are tuned with intelligent grid search (IGS) method for the derived manifold coordinates to achieve optimal classification accuracies. Of particular interest is its performance with small training size, because labelled samples are usually scarce due to its high acquisition cost. The proposed spatial kernel PCA (KPCA) is compared with PCA in terms of classification accuracy with the nearest-neighbourhood classification method.

• Journal of applied mathematics & informatics
• /
• v.39 no.5_6
• /
• pp.851-858
• /
• 2021

In general, secondary batteries are widely used as an electric energy source. Among them, the state of energy storage of mobile devices is very important information. As a method of estimating a state, there is a method of estimating the state by integrating the current according to an energy storage state of a battery, and a method of designing a state estimator by measuring a voltage and estimating a charge amount based on a battery model. In this study, we designed the state estimator using an extended Kalman filter to increase the precision of the state estimation of the charge amount by including the error of the system model and having the robustness to the noise.

• Journal of the Korean Society of Industry Convergence
• /
• v.22 no.3
• /
• pp.265-272
• /
• 2019

For tracking a varying number of people using laser range finder, it is important to deal with appearance/disappearance of people due to various causes including occlusions. We propose a method for tracking people with automatic initialization by integrating observations from laser range finder. In our method, the problem of estimating 2D positions and orientations of multiple people's walking direction is formulated based on a mixture kalman filter. Proposal distributions of a kalman filter are constructed by using a mixture model that incorporates information from a laser range scanner. Our experimental results demonstrate the effectiveness and robustness of our method.

• Journal of Korean Dental Science
• /
• v.15 no.2
• /
• pp.166-171
• /
• 2022

This case report describes a dynamic digital esthetic rehabilitation procedure that integrates a new three-dimensional augmented reality (3D-AR) technique to treat a patient with multiple missing anterior teeth. The prostheses were designed using computer-aided design (CAD) software and virtually trialed using static and dynamic visualization methods. In the static method, the prostheses were visualized by integrating the CAD model with a 3D face scan of the patient. For the dynamic method, the 3D-AR application was used for real-time tracking and projection of the CAD prostheses in the patient's mouth. Results of a quick survey on patient satisfaction with the two visualization methods showed that the patient felt more satisfied with the dynamic visualization method because it allowed him to observe the prostheses directly on his face and be more proactive in the treatment process.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.253-260
• /
• 2022

Existing web-based cost databases have proved invaluable for construction cost estimating. These databases have been utilized to compute approximate cost estimates using assembly rates, unit rates, and etc. These web-based databases can be used independently with traditional cost estimation methods (manual methods) or used to support BIM-based cost estimating platforms. However, these databases are rigid, costly, and require a lot of manual inputs to reflect recent trends in prices or prices relative to a construction project's location. To address this gap, this study integrated deep learning techniques with web-based price analysis to develop a database that incorporates a project's location cost estimating standards and current cost trends in generating a cost estimate. The proposed method was tested in a case study project in Lagos, Nigeria. A cost estimate was successfully generated. Comparison of the experimental results with results using current industry standards showed that the proposed method achieved a 98.16% accuracy. The results showed that the proposed method was successful in generating approximate cost estimates irrespective of project's location.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.6
• /
• pp.1620-1634
• /
• 2023

This paper studies the task of embedding nodes with multiple graphs representing multiple information channels, which is useful in a large volume of network clustering tasks. By learning a node using multiple graphs, various characteristics of the node can be represented and embedded stably. Existing studies using multi-channel networks have been conducted by integrating heterogeneous graphs or limiting common nodes appearing in multiple graphs to have similar embeddings. Although these methods effectively represent nodes, it also has limitations by assuming that all networks provide the same amount of information. This paper proposes a method to overcome these limitations; The proposed method gives different weights according to the source graph when embedding nodes; the characteristics of the graph with more important information can be reflected more in the node. To this end, a novel method incorporating a multi-channel gate layer is proposed to weigh more important channels and ignore unnecessary data to embed a node with multiple graphs. Empirical experiments demonstrate the effectiveness of the proposed multi-channel-based embedding methods.