• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.26-38
• /
• 2021

In this paper, DLKF (Deep Learning Keypoint Filtering), the deep learning-based keypoint filtering method for the rapidization of the image registration method for remote sensing images is proposed. The complexity of the conventional feature-based image registration method arises during the feature matching step. To reduce this complexity, this paper proposes to filter only the keypoints detected in the artificial structure among the keypoints detected in the keypoint detector by ensuring that the feature matching is matched with the keypoints detected in the artificial structure of the image. For reducing the number of keypoints points as preserving essential keypoints, we preserve keypoints adjacent to the boundaries of the artificial structure, and use reduced images, and crop image patches overlapping to eliminate noise from the patch boundary as a result of the image segmentation method. the proposed method improves the speed and accuracy of registration. To verify the performance of DLKF, the speed and accuracy of the conventional keypoints extraction method were compared using the remote sensing image of KOMPSAT-3 satellite. Based on the SIFT-based registration method, which is commonly used in households, the SURF-based registration method, which improved the speed of the SIFT method, improved the speed by 2.6 times while reducing the number of keypoints by about 18%, but the accuracy decreased from 3.42 to 5.43. Became. However, when the proposed method, DLKF, was used, the number of keypoints was reduced by about 82%, improving the speed by about 20.5 times, while reducing the accuracy to 4.51.

• Journal of the Society of Disaster Information
• /
• v.18 no.4
• /
• pp.883-890
• /
• 2022

Purpose: Currently, in the case of domestic fire fighting facility design, it is difficult to secure highquality manpower due to low design costs and overheated competition between companies, so there is a limit to improving the fire safety performance of buildings. Accordingly, AI-based firefighting design solutions were studied to solve these problems and secure leading fire engineering technologies. Method: Through AutoCAD, which is widely used in existing fire fighting design, the procedures required for basic design and implementation design were processed, and AI technology was utilized through the YOLO v4 object recognition deep learning model. Result: Through the design process for fire fighting facilities, the facility was determined and the drawing design automation was carried out. In addition, by learning images of doors and pillars, artificial intelligence recognized the part and implemented the function of selecting boundary areas and installing piping and fire fighting facilities. Conclusion: Based on artificial intelligence technology, it was confirmed that human and material resources could be reduced when creating basic and implementation design drawings for building fire protection facilities, and technology was secured in artificial intelligence-based fire fighting design through prior technology development.

• Journal of Internet Computing and Services
• /
• v.24 no.4
• /
• pp.25-36
• /
• 2023

Today, as AI (Artificial Intelligence) technology is introduced in various fields, including security, the development of technology is accelerating. However, with the development of AI technology, attack techniques that cleverly bypass malicious behavior detection are also developing. In the classification process of AI models, an Adversarial attack has emerged that induces misclassification and a decrease in reliability through fine adjustment of input values. The attacks that will appear in the future are not new attacks created by an attacker but rather a method of avoiding the detection system by slightly modifying existing attacks, such as Adversarial attacks. Developing a robust model that can respond to these malware variants is necessary. In this paper, we propose two methods of generating Adversarial attacks as efficient Adversarial attack generation techniques for improving Robustness in AI models. The proposed technique is the XAI-based attack technique using the XAI technique and the Reference based attack through the model's decision boundary search. After that, a classification model was constructed through a malicious code dataset to compare performance with the PGD attack, one of the existing Adversarial attacks. In terms of generation speed, XAI-based attack, and reference-based attack take 0.35 seconds and 0.47 seconds, respectively, compared to the existing PGD attack, which takes 20 minutes, showing a very high speed, especially in the case of reference-based attack, 97.7%, which is higher than the existing PGD attack's generation rate of 75.5%. Therefore, the proposed technique enables more efficient Adversarial attacks and is expected to contribute to research to build a robust AI model in the future.

• Journal of Applied Reliability
• /
• v.15 no.4
• /
• pp.262-269
• /
• 2015

Accelerated degradation mechanism of the heat-resistant steel for high temperature plant was analysed in terms of microstructure and hardness. In order to simulate the microstructure of the steel actually used at $540^{\circ}C$ in the field, isothermal exposure was carried out at $630^{\circ}C$ up to 4,800 hours. The artificial degradation mechanism was comparatively verified to successfully simulate degradation of the long-time used field material. For the artificially degraded specimens, databases including size and aspect ratio of carbide, chemical composition (i.e., Cr/Mo ratio) of grain boundary carbide were built up. These degradation parameters were suggested as fingerprints for PHM (i.e., prognostics health management) of power plants.

• Journal of the Korea Computer Graphics Society
• /
• v.4 no.1
• /
• pp.55-66
• /
• 1998

This paper presents a simple method for relieving the ambiguity problems within the sub-voxel based surface-fitting approach for the surface construction. ECB algorithm is proposed to avoid the ambiguity problem which is the root of the holes within the resulting polygon based approximation. The basic idea of our disambiguation strategy is the use of a set of predefined modeling primitives (we call SMP) which guarantees the topological consistency of resulted surface polygons. 20 SMPs are derived from the extension of the concept of the elementary modeling primitives in the CB algorithm [3], and fit one to five faces of them to the iso-surface crossing a cell with no further processing. A look-up table which has a surface triangle list is pre-calculated using these 20 SMPs. All of surface triangles in the table are from the faces of SMPs and are stored in the form of edge list on which vertices of each surface triangle are located. The resulted polygon based approximation is unique at every threshold value and its validity is guaranteed without considering the complicated problems such as average of density and postprocessing. ECB algorithm could be free from the need for the time consuming post-processing, which eliminates holes by revisiting every boundary cell. Through three experiments of surface construction from volume data, its capability of hole avoidance is showed.

• Journal of KSNVE
• /
• v.8 no.2
• /
• pp.239-250
• /
• 1998

Traditionally, intake noise from internal combustion engine has not recevied much attention compared to exhaust noise. But nowadays, intake noise is a major contributing factor to automotive passenger compartment noise levels. The main objective of this paper is to identify the mechanism of generation, propagation and radiation of the intake noise. With a simplest geometric model, one of the main noise sources for the intake stroke is found to be the pressure surge, which is generated after intake valve closing. The pressure surge, which has the nonlinear acoustic behavior, propagates and radiates with relatively large amplitude. In this paper, unsteady compressible Navier-Stokes equations are employed for the intake stroke of axisymmetric model having a single moving cylinder and a single moving intake valve. To simulate the periodic motion of the piston and the valve, unsteady deforming mesh algorithm is employed and Thompson's non-reflecting boundary condition is applied to the radiation field. In order to resolve the small amplitude waves at the radiation field, essentially non-oscillatory(ENO) schemes with an artificial compression method (ACM) are used.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.1E
• /
• pp.32-35
• /
• 2006

The screech tone of underexpanded jet is numerically calculated without any specific modeling for the screech tone itself. Fourth-order optimized compact scheme and fourth-order Runge-Kutta method are used to solve the 2D axisymmetric Euler equation. Adaptive nonlinear artificial dissipation model and generalized characteristic boundary condition are also used. The screech tone, generated by a closed loop between instability waves and quasi-periodic shock cells at the near field, is reasonably analyzed with present numerical methods for the underexpanded jet having Mach number 1.13. First of all, the centerline mean pressure distribution is calculated and compared with experimental and other numerical results. The instantaneous density contour plot shows Mach waves due to mixing layer convecting supersonically, which propagate downstream. The pressure signal and its Fourier transform at upstream and downstream shows the directivity pattern of screech tone very clearly. Most of all, we can simulate the axisymmetric mode change of screech tone very precisely with present method. It can be concluded that the basic phenomenon of screech tone including the frequency can be calculated by using high-order and high-resolution schemes without any specific numerical modeling for screech tone feedback loop.

• The Transactions of the Korean Institute of Electrical Engineers C
• /
• v.48 no.6
• /
• pp.489-493
• /
• 1999

In this paper, an analysis of microstrip-slotline transition is performed using a 3D vector Finite Element Method(FEM). Artificial anistropic absorber technique is employed to implement an matching boundary condition in FEM. On the base of the analysis, power divider/combiner is designed. The structure of the power combiner already developed are Branch-line coupler, Rat-race coupler, Wilkinson coupler, Lange coupler, etc. Which are all planar, If the frequency goes up, the coupling efficiency of these planar couplers is decreased on account of skin loss. Especially, in millimeter-wave band, the efficiency of more than two ways combiner is radically reduced, so that application in power amplifier circuit is almost impossible, Microstrip-slotline transition structure is a power combining technique integrated into wave-guide, so that the loss is small and the efficiency is high. Theoretically, we can mount several transistors into the power-combiner. This makes it possible to develop a high power amplifier. The numerically calculated performances of the device that is, we believe, the best are compared to the experimental results in Ka-Band(26.5GHz-40GHz).

• Journal of Ocean Engineering and Technology
• /
• v.34 no.4
• /
• pp.245-252
• /
• 2020

Wave run-up is an important phenomenon that should be considered in ocean structure design. In this study, the wave run-up of a surface-piercing circular cylinder was calculated in the time domain using the three-dimensional linear and fully nonlinear numerical wave tank (NWT) techniques. The NWT was based on the boundary element method and the mixed Eulerian and Lagrangian method. Stokes second-order waves were applied to evaluate the effect of the nonlinear waves on wave run-up, and an artificial damping zone was adopted to reduce the amount of reflected and re-reflected waves from the sidewall of the NWT. Parametric studies were conducted to determine the effect of wavelength, wave steepness, and the draft of the cylinder on the wave run-up of the cylinder. The maximum wave run-up value occurred at 0°, which was in front of the cylinder, and the minimum value occurred near the circumferential angle of 135°. As the diffraction parameter increased, the wave run-up increased up to 1.7 times the wave height. Furthermore, the wave run-up was 4% higher than the linear wave when the wave steepness was 1/35. In particular, the crest height of the wave run-up increased by 8%.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2006.06a
• /
• pp.327-335
• /
• 2006

Case Based Reasoning (CBR) has come to be considered as an appropriate technique for diagnosis, prognosis and prescription in medicine. However, canventional CBR has a limitation in that it cannot incorporate asymmetric misclassification cast. It assumes that the cast of type1 error and type2 error are the same, so it cannot be modified according ta the error cast of each type. This problem provides major disincentive to apply conventional CBR ta many real world cases that have different casts associated with different types of error. Medical diagnosis is an important example. In this paper we suggest the new knowledge extraction technique called Cast-Sensitive Case Based Reasoning (CSCBR) that can incorporate unequal misclassification cast. The main idea involves a dynamic adaptation of the optimal classification boundary paint and the number of neighbors that minimize the tatol misclassification cast according ta the error casts. Our technique uses a genetic algorithm (GA) for finding these two feature vectors of CSCBR. We apply this new method ta diabetes datasets and compare the results with those of the cast-sensitive methods, C5.0 and CART. The results of this paper shaw that the proposed technique outperforms other methods and overcomes the limitation of conventional CBR.