• Journal of Korea Multimedia Society
• /
• v.25 no.8
• /
• pp.1212-1223
• /
• 2022

Skin lesions are common diseases that range from skin rashes to skin cancer, which can lead to death. Note that early diagnosis of skin diseases can be important because early diagnosis of skin diseases considerably can reduce the course of treatment and the harmful effect of the disease. Recently, the development of computer-aided diagnosis (CAD) systems based on artificial intelligence has been actively made for the early diagnosis of skin diseases. In a typical CAD system, the accurate classification of skin lesion types is of great importance for improving the diagnosis performance. Motivated by this, we propose a novel deep ensemble classification with multi-scale attention networks. The proposed deep ensemble networks are jointly trained using a single loss function in an end-to-end manner. In addition, the proposed deep ensemble network is equipped with a multi-scale attention mechanism and segmentation information of the original skin input image, which improves the classification performance. To demonstrate our method, the publicly available human skin disease dataset (HAM 10000) and the private animal skin lesion dataset were used for the evaluation. Experiment results showed that the proposed methods can achieve 97.8% and 81% accuracy on each HAM10000 and animal skin lesion dataset. This research work would be useful for developing a more reliable CAD system which helps doctors early diagnose skin diseases.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.38 no.7
• /
• pp.727-734
• /
• 2014

This paper describes multi-scale based ductile fracture simulation using finite element (FE) damage analysis. The maximum and crack initiation loads of cracked components were predicted using proposed virtual testing method. To apply the local approach criteria for ductile fracture, stress-modified fracture strain model was adopted as the damage criteria with modified calibration technique that only requires tensile and fracture toughness test data. Element-size-dependent critical damage model is also introduced to apply the proposed ductile fracture simulation to large-scale components. The results of the simulation were compared with those of the tests on SA333 Gr. 6 full-scale pipes at $288^{\circ}C$, performed by the Battelle Memorial Institute.

• Korean Journal of Remote Sensing
• /
• v.34 no.5
• /
• pp.739-747
• /
• 2018

Integration analysis of multi-sensor satellite images is becoming increasingly important. The first step in integration analysis is image registration between multi-sensor. SIFT (Scale Invariant Feature Transform) is a representative image registration method. However, optical image and SAR (Synthetic Aperture Radar) images are different from sensor attitude and radiation characteristics during acquisition, making it difficult to apply the conventional method, such as SIFT, because the radiometric characteristics between images are nonlinear. To overcome this limitation, we proposed a modified method that combines the SAR-SIFT method and shape descriptor vector DLSS(Dense Local Self-Similarity). We conducted an experiment using two pairs of Cosmo-SkyMed and KOMPSAT-2 images collected over Daejeon, Korea, an area with a high density of buildings. The proposed method extracted the correct matching points when compared to conventional methods, such as SIFT and SAR-SIFT. The method also gave quantitatively reasonable results for RMSE of 1.66m and 2.45m over the two pairs of images.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.14 no.3
• /
• pp.115-122
• /
• 2004

When porous materials are dried, the particles flocculate into fish-net structure in gel phase. In order to exactly analyze the stress distribution of porous materials during drying process, the elastic tensor of microscopic gel structures has to be predicted considering pore shapes as well as porosities of porous materials. The elastic characteristics of porous materials associated with porosities were predicted analyzing microscopic gel structures with circular and cross pores via homogenization method and the drying processes of the electric porous ceramic insulator were simulated using finite element method (FEM). Comparing analysis results between consideration and negligence of pores, the deformed shape and distributions of temperature and moisture were similar but the residual stress was significantly different.

• Structural Engineering and Mechanics
• /
• v.63 no.6
• /
• pp.761-769
• /
• 2017

Accurately extracting the axle distribution information of a passing vehicle from bridge dynamic responses experiences a key and challenging step in non-pavement bridge weigh-in-motion (BWIM). In this article, the wavelet transformation is adopted and the wavelet coefficient curve is used as a substitute for dynamic response. The driving frequency is introduced and expanded to multi-axle vehicle, and the wavelet coefficient curve on specific scale corresponding to the driving frequency is confirmed to contain obvious axle information. On this basis, an automatic method for axle distribution information identification is proposed. The specific wavelet scale can be obtained through iterative computing, and the false peaks due to bridge vibration can be eliminated through cross-correlation analysis of the wavelet coefficients of two measure points. The integrand function that corresponds to the maximum value of the cross-correlation function is used to identify the peaks caused by axles. A numerical application of the proposed axle information identification method is carried out. Numerical results demonstrate that this method acquires precise axle information from the responses of an axle-insensitive structure (e.g., girder) and decreases the requirement of sensitivity structure of BWIM. Finally, an experimental study on a full-scale simply supported bridge is also conducted to verify the effectiveness of this method.

• Transactions of Materials Processing
• /
• v.17 no.8
• /
• pp.627-632
• /
• 2008

The cross sectional shape of intermediate die is one of important parameters to improve dimensional accuracy of final product in shaped drawing process. Until now, it has been designed by the experience or trial and error of the expert. In this study, the cross sectional shape of intermediate die for spline shape is determined by the electric fields analysis and scale factor method. The result of the electric fields analysis and scale factor method have been compared with that of the expert method. The effects of cross sectional shape on the dimensional accuracy were investigated by using FE-simulation. And then the multi-stage shaped drawing experiments were performed to verify the results of FE-simulation. As a result, the cross sectional shape from the electric fields analysis and scale factor method had the good dimensional accuracy. These two methods can be used for the method to obtain the cross sectional shape of intermediate die in shaped drawing process.

• Proceedings of the KSME Conference
• /
• 2004.11a
• /
• pp.400-403
• /
• 2004

We present a coarse-graining model to describe the mechanical behaviors of multi-walled carbon nanotubes. To find the atomic configuration in membrane-like nanostructure i.e. carbon nanotube, we employ interpolation functions and the associated element-variables that are defined in the subdivided region. Tersoff-Brenner potential is adopted for interaction of bonded atoms and also van der Waals force for non-bonded interaction. Moreover, we simulate the coarse-graining multi-walled carbon nanotubes with defects and its result is compared with that of perfect multi-walled carbon nanotubes.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.29 no.5
• /
• pp.449-457
• /
• 2011

As automated image processing techniques have been required in multi-temporal/multi-sensor geospatial image applications, use of automated but highly invariant image matching technique has been a critical ingredient. Note that there is high possibility of geometric and spectral differences between multi-temporal/multi-sensor geospatial images due to differences in sensor, acquisition geometry, season, and weather, etc. Among many image matching techniques, the SIFT (Scale Invariant Feature Transform) is a popular method since it has been recognized to be very robust to diverse imaging conditions. Therefore, the SIFT has high potential for the geospatial image processing. This paper presents a performance test results of the SIFT on geospatial imagery by simulating various image differences such as shear, scale, rotation, intensity, noise, and spectral differences. Since a geospatial image application often requires a number of good matching points over the images, the number of matching points was analyzed with its matching positional accuracy. The test results show that the SIFT is highly invariant but could not overcome significant image differences. In addition, it guarantees no outlier-free matching such that it is highly recommended to use outlier removal techniques such as RANSAC (RANdom SAmple Consensus).

• Journal of IKEEE
• /
• v.22 no.3
• /
• pp.810-815
• /
• 2018

The kernelized correlation filter algorithm yielded meaningful results in accuracy for object tracking. However, because of the use of a fixed size template, we could not cope with the scale change of the tracking object. In this paper, we propose a method to track objects by finding the best scale for each frame using correlation filtering response values in multi-scale using nearest neighbor interpolation and Gaussian normalization. The scale values of the next frame are updated using the optimal scale value of the previous frame and the optimal scale value of the next frame is found again. For the accuracy comparison, the validity of the proposed method is verified by using the VOT2014 data used in the existing kernelized correlation filter algorithm.

• Composites Research
• /
• v.31 no.5
• /
• pp.251-259
• /
• 2018

In this paper, statistical volume element modeling method was developed for multi-scale progressive failure analysis of fiber reinforced composite materials. Big-size statistical volume elements (BSVEs) was considered to minimize the size effect in the micro-scale, by including as many fibers as possible. For that purpose, a mesh cutting method is suggested and adapted into the fiber model generator that creates finite element domain rapidly. The fiber defect model was also developed based on the experimental distribution of the fiber strength. The size effects from the local load sharing (LLS) are evaluated by increasing the fiber inclusion in the micro-scale model. Finally, continuum damage mechanics (CDM) model to the fiber direction was extracted from numerical analysis on BSVEs. And it was compared with strength prediction from typical representative volume element (RVE) model.