• Journal of the Korea Convergence Society
• /
• v.11 no.12
• /
• pp.39-47
• /
• 2020

In this paper, we proposed a deep learning-based detection method using Saliency Attention to detect salient objects in images. The salient object detection separates the object where the human eye is focused from the background, and determines the highly relevant part of the image. It is usefully used in various fields such as object tracking, detection, and recognition. Existing deep learning-based methods are mostly Autoencoder structures, and many feature losses occur in encoders that compress and extract features and decoders that decompress and extend the extracted features. These losses cause the salient object area to be lost or detect the background as an object. In the proposed method, Saliency Attention is proposed to reduce the feature loss and suppress the background region in the Autoencoder structure. The influence of the feature values was determined using the ELU activation function, and Attention was performed on the feature values in the normalized negative and positive regions, respectively. Through this Attention method, the background area was suppressed and the projected object area was emphasized. Experimental results showed improved detection results compared to existing deep learning methods.

• Journal of Information Technology Services
• /
• v.12 no.1
• /
• pp.309-321
• /
• 2013

In current complex nature of management with task-structures, a method to reach the enterprise's informationalization success is not common. To satisfy these various requirement, improving the usability of information technology (IT) is a key factor which defining the level of organizational requirement first. Imposing an IT-solution which has excess service of the organization's previous task-environment, procedure and scope is not effective to SME-level unit, which unit could not have a formal organization structure and task structure. SME level informationalization will be success if each function realizes easier on the task-employee's viewpoint. Achieving this objective, a solution provider or department must reflect their work characteristics of nature which has least level of work performing resistance. It is most useful system for SME level unit, if a provider develops single programs which based on task activities, and each program can configure network-linking.

• The KIPS Transactions:PartD
• /
• v.9D no.5
• /
• pp.827-836
• /
• 2002

Moving object management systems manage spatiotemporal data, which change their location over tine such as people, animals, and cars. These moving object management systems can be applied to vehicle location tracking, digital battlefield, location-based service, and so on. The existing moving object management systems only manage past or future location of the moving objects separately. Therefore, they cannot suggest estimation method of uncertain past or future location of the moving objects. In this paper, we propose a moving object management system, which not only manages historical data of the moving objects, but also predicts past and future location of the moving objects using historical data stored in database. We define the moving objects for vehicle location tracking and propose a moving object database structure. Finally, we suggest an execution model of the proposed system and apply the execution model to a virtual scenario for vehicle tracking.

• Journal of Korea Game Society
• /
• v.9 no.4
• /
• pp.97-106
• /
• 2009

In this paper, a realtime approach to the simulation of virtual paper with an adaptive mesh structure is proposed. The proposed method can be applied to arbitrary triangular mesh structures and efficiently produces wrinkles and creases on the paper surface with stable numerical integration and deformation-based mesh refinement. In order to plausibly represent the crumples on the paper object, we employed an adaptive mesh structure with breakable springs. Because the adaptive structure continuously inserts or removes vertices and edges to or from the mesh structure, the conservation of the mass and the momentum should be carefully taken into account for the plausible simulation of the virtual paper. The proposed method produced plausible animation of paper-like thin shell in realtime environments.

• Journal of Information Technology Services
• /
• v.11 no.4
• /
• pp.277-294
• /
• 2012

The purpose of this research is to analyze big data structure and various objects in big data industry based on ecological perspective. Big data is rapidly emerging as a highly valuable resource to secure competitiveness of enterprise and government. Accordingly, the main issues in big data are to find ways of creating economic value and solving various problems. However big data is not systematically organized, and hard to utilize as it constantly expands to related industry such as telecommunications, finance and manufacturing. Under this circumstance, it is crucial to understand range of big data industry and to which stakeholders are related. The ecological approach is useful to understand comprehensive industry structure. Therefore this study aims at confirming big data structure and finding issues from interaction among objects. Results of this study show main framework of big data ecosystem including relationship among object elements composing of the ecosystem. This study has significance as an initial study on big data ecosystem. The results of the study can be useful guidelines to the government for making systemized big data ecosystem and the entrepreneur who is considering launching big data business.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2006.04a
• /
• pp.223-228
• /
• 2006

Monitoring structures is important to maintain the safety and serviceability of the structures. The maximum displacement in the structure should be precisely and frequently monitored because it is a direct assessment index indicating its stiffness. However, no practical method has been developed to monitor such displacement precisely, particularly for high-rise buildings and long span bridges because they cannot be easily accessible. To overcome such difficult accessibility, we propose to use a LIDAR system that remotely samples the surface of an object using laser pulses and generates the coordinates of numerous points on the surface. By analyzing the LIDAR points sampled from the surfaces of a deformed structure, we can precisely determine the displacement of the structure. In this study, we thus develop a novel method based the LIDAR system and perform an indoor experiment to prove its performance. This experimental results strongly supports that the displacement measurement using the LIDAR system are enough accurate to be used for structural analyses.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.8
• /
• pp.1227-1234
• /
• 2001

A multi d.o.f robotic manipulator is considered for multi-axis vibration control of a slender structure, using the concept of the flow source based vibration control. In order not to cause the motion saturation of the manipulator system, a hybrid dynamics associated with the flexible and desired manipulator error dynamics is also modeled as the control object. It is numerically shown that the flexible vibrations and the base motions of a test structure can be effectively controlled with the proposed hybrid dynamics.

• English Language & Literature Teaching
• /
• v.10 no.1
• /
• pp.179-200
• /
• 2004

The purpose of this study is to investigate the adequacy of derivational approach and representational approach to syntactic theories in generative grammar. As the generative grammar is based on the derivational process of syntactic theories, it is suggested that derivational approach is more valid than representational approach. Move, Economy Principle, Local Economy and Label-free Phrase Structure in Minimalist framework support the preference of derivational approach to representational approach with the elimination of computational complexity, minimality, and label-free phrase structure. Syntactic structure is considered as the result of the interaction of the properties of lexical items containing probe and economy conditions constrained by bare output conditions. On the other hand, Pseudogapping in Lasnik(1999) is analyzed in terms of Object Shift, that is, overt raising to Spec of $Agr_o$ and the PF deletion of VP in representational approach. Hence, it is suggested that the combination of derivational and representational approaches to syntactic theories can be admitted in generative grammar.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.8
• /
• pp.2473-2489
• /
• 2022

Urban development has brought about the increasing saturation of urban traffic demand, and traffic congestion has become the primary problem in transportation. Roads are in a state of waiting in line or even congestion, which seriously affects people's enthusiasm and efficiency of travel. This paper mainly studies the discrete domain path planning method based on the flow data. Taking the traffic flow data based on the highway network structure as the research object, this paper uses the deep learning theory technology to complete the path weight determination process, optimizes the path planning algorithm, realizes the vehicle path planning application for the expressway, and carries on the deployment operation in the highway company. The path topology is constructed to transform the actual road information into abstract space that the machine can understand. An appropriate data structure is used for storage, and a path topology based on the modeling background of expressway is constructed to realize the mutual mapping between the two. Experiments show that the proposed method can further reduce the interpolation error, and the interpolation error in the case of random missing is smaller than that in the other two missing modes. In order to improve the real-time performance of vehicle path planning, the association features are selected, the path weights are calculated comprehensively, and the traditional path planning algorithm structure is optimized. It is of great significance for the sustainable development of cities.

• Journal of the Korea Computer Graphics Society
• /
• v.18 no.1
• /
• pp.29-34
• /
• 2012

High-quality tetrahedral meshes are crucial for FEM-based simulation of large elasto-plastic deformation and tetrahedral-mesh-based simulation of fluid flow. This paper proposes a volume meshing method that exploits an octree-based adaptive signed distance field to fill the inside of a polygonal object with tetrahedra, of which dihedral angles are good. The suggested method utilizes an octree structure to reduce the total number of tetrahedra by space-efficiently filling an object with graded tetrahedra. To obtain a high-quality mesh with good dihedral angles, we restrict the octree in such a way that any pair of neighboring cells only differs by one level. In octree-based tetrahedral meshing, the signed distance computation of a point to the surface of a given object is a very important and frequently-called operation. To accelerate this operation, we develop a method that computes a signed distance field directly on the vertices of the octree cells while constructing the octree using a top-down approach. This is the main focus of the paper. The suggested tetrahedral meshing method is fast, stable and easy to implement.