• Journal of the Korea Society of Computer and Information
• /
• v.27 no.11
• /
• pp.29-38
• /
• 2022

This paper proposes a system for classifying gait types using an ensemble deep learning network for gait data measured by a smart insole equipped with multi-sensors. The gait type classification system consists of a part for normalizing the data measured by the insole, a part for extracting gait features using a deep learning network, and a part for classifying the gait type by inputting the extracted features. Two kinds of gait feature maps were extracted by independently learning networks based on CNNs and LSTMs with different characteristics. The final ensemble network classification results were obtained by combining the classification results. For the seven types of gait for adults in their 20s and 30s: walking, running, fast walking, going up and down stairs, and going up and down hills, multi-sensor data was classified into a proposed ensemble network. As a result, it was confirmed that the classification rate was higher than 90%.

• Journal of Broadcast Engineering
• /
• v.27 no.4
• /
• pp.487-498
• /
• 2022

Depth map is an image that contains distance information in 3D space on a 2D plane and is used in various 3D vision tasks. Many existing depth estimation studies mainly use narrow FoV images, in which a significant portion of the entire scene is lost. In this paper, we propose a technique for generating 360° omnidirectional RGBD images from a sparse set of narrow FoV images. The proposed generative adversarial network based image generation model estimates the relative FoV for the entire panoramic image from a small number of non-overlapping images and produces a 360° RGB and depth image simultaneously. In addition, it shows improved performance by configuring a network reflecting the spherical characteristics of the 360° image.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.2A
• /
• pp.215-222
• /
• 2008

This paper introduces an experimental verification of a tension estimation method based on system identification approach for a double hanger system on a suspension bridge. A laboratory model of such double hanger system has been made for this study. Total nine cases of the vibration tests have been conducted with respect to three levels of applied tension and three cases of the location of clamp. For a set of the collected acceleration response data, modal analysis has been followed in order to extract the natural frequencies and mode shapes of the selected cable systems. For the extracted modal parameters, the existing tension estimation methods based on the string theory and axially loaded beam theory have been firstly applied to estimate the tensile force on the double hanger cable system. Next, the tensile force on cables has been estimated by the system identification approach. It is seen that the errors in the tension estimation using the frequency-based system identification technique are about 3% for all cases while the estimation error using the existing method is up to 53.1%.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.317-326
• /
• 2024

Biometric recognition is a technology that determines whether a person is identified by extracting information on a person's biometric and behavioral characteristics with a specific device. Cyber threats such as forgery, duplication, and hacking of biometric characteristics are increasing in the field of biometrics. In response, the security system is strengthened and complex, and it is becoming difficult for individuals to use. To this end, multiple biometric models are being studied. Existing studies have suggested feature fusion methods, but comparisons between feature fusion methods are insufficient. Therefore, in this paper, we compared and evaluated the fusion method of multiple biometric models using fingerprint, face, and iris images. VGG-16, ResNet-50, EfficientNet-B1, EfficientNet-B4, EfficientNet-B7, and Inception-v3 were used for feature extraction, and the fusion methods of 'Sensor-Level', 'Feature-Level', 'Score-Level', and 'Rank-Level' were compared and evaluated for feature fusion. As a result of the comparative evaluation, the EfficientNet-B7 model showed 98.51% accuracy and high stability in the 'Feature-Level' fusion method. However, because the EfficietnNet-B7 model is large in size, model lightweight studies are needed for biocharacteristic fusion.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.3
• /
• pp.209-214
• /
• 2017

Most of the present potential transformers of train vehicles are of the oil-type filled with insulating oil and are susceptible to problems such as explosion due to the increase in the internal pressure during train operation and poor reliability near the end of their life cycle. As a solution to this problem, it is necessary to develop a molded dry-type potential transformer with excellent pressure-resistance performance using insulating resin. In order to localize the product, the Korea Railroad Research Institute has been developing a molded dry-type potential transformer. As part of this research, it is necessary to analyze the vibration characteristics of the developed product and to check the transformer performance in a vibration environment. In this study, a resonance test and simulated long-term life test of the developed product were conducted according to the KS R 9144 and IEC 61373 standards, respectively, which are vibration test methods for railway vehicle parts. Their natural frequencies were analyzed by comparing the results of the numerical modal analysis and resonance test, in order to confirm their adherence to the standards. Also, the performance test after the simulated long-term life test confirmed that the operation of the developed transformer was not problematic even in a long-time vibration environment.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 1993.04a
• /
• pp.35-39
• /
• 1993

항공기의 플러터 현상을 예방키 위한 개발시의 주요 공학적 임무로는 플러 터 해석, 풍동 플러터 시험, 실기체에 대한 지상진동시험 및 비행 플러터 시 험이 있다. 이들 업무는 군용 항공기의 경우에는 군사 규격서, 민간 항공기 의 경우는 FAR 규정 등에 항공기 개발 및 개조시에는 필히 수행토록 명시 되어 있으며, 특히 개발항공기의 인증을 받기 위해서는 초도 비행전까지 필 수적으로 완료되어야 하는 업무이다. 이중 항공기 지상진동시험은 개발항공 기의 초도 비행전에 실기체를 대상으로 하는 구조 동특성 규명시험으로써 플러터 해석 모델의 정확도를 입증하고 그 해석모델을 수정하거나 개선시키 는데 필요한 구조의 동특성 변수들을 실험적으로 규명하는 시험이다. 이 시 험은 개발된 항공기의 초도비행허가를 획득하고 비행속도제한을 설정키 위 해서 초도비행 직전에 초도 개발 비행기를 대상으로 필수적으로 수행되어야 하는 필수적 시험이다. 이에 따라 개발된 항공기의 전기체 지상진동시험을 수행하였고, 시험데이터의 모달해석을 하였다. 이 시험을 수행하기 위해 3개 의 가진기를 사용하였고, 모두 159개의 지점에서 주파수응답 함수를 얻었다. 최대 48 채널의 데이터 획득시스템을 사용하였으므로, 네번에 걸쳐 측정위치 를 옮겨 전체데이타를 획득하였다. 지상진동 시험의 최종해석 결과는 유한요 소 모델의 유효성을 검토하는데 사용되었고, 시험 데이타를 이용한 가진응답 해석(Forced Response Analysis)을 통하여 비행플러터 시험시에 사용되는 플러터 가진기의 위치선정 및 가진력의 크기를 결정하는데 이용하였다.ncy)가 저주 파수대로 천이(Lower Shift)하는 현상이 나타났으며, 일정한 작업량이라도 작업중량을 줄이고 작업 빈도를 높여서 작업할 경우에 이러한 현상이 더욱 두드러졌다. 작업시간의 경과에 따른 MVIS의 감소 현상은 작업빈도가 높고 비대칭형 작업일 때 가장 크게 나타났다. 총손실을 줄이고, 상대적 비용절감효과를 갖게할 수 있다고 하였다. 따라서 본 연구에서는 성인 여성기성복의 치수적합성을 높이기위하여 출산 후 중년에 접어 들면서 체형이 변화되는 것을 고려하여 연령을 분류하고(18세-34세, 30세-51세), 각 연령 집단에 따른 체형을 각각 3가지로 분류하였다. 이에 따라 의복 생산시의 총손실을 줄이기위한 상의, 하의생산시 필요한 부위별 최적규격치 간격을 제시하였다.적인 기능으로 보여지는 것에 좁혀서 작업능력의 연령증가 변화에 대하여다원적 평가를 하는 것이 실제적이라고 할 수 있다. 따라서 본 연구에서는 인간이 가지고 있는 다수의 기능중에서 수지교 치성과 연령증가와의 관계를 조사한다. 만약 연령증가 만으로 수지교치성을 평가 할 수 없는 경우에는 어떠한 요인이 수지기민성의 변화에 영향을 미치는가를 검토한다.t list)에서 자동적으로 사건들의 순서가 결정되도록 확장하였으며, 설비 제어방식에 있어서도 FIFO, LIFO, 우선 순위 방식등을 선택할 수 있도록 확장하였다. SIMPLE는 자료구조 및 프로그램이 공개되어 있으므로 프로그래머가 원하는 기능을 쉽게 추가할 수 있는 장점도 있다. 아울러 SMPLE에서 새로이 추가된 자료구조와 함수 및 설비제어 방식등을 활용하여 실제 중형급 시스

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.26 no.3
• /
• pp.270-280
• /
• 2016

This paper investigates the evolution of EM admittance of piezoelectric transducers mounted on a notched beam from wave propagation perspective. A finite element analysis is adopted to obtain numerical solutions for Lamb waves reverberating on the notched beam. The mode-converted Lamb wave signals due to a notch are extracted by using the polarization characteristics of piezoelectric transducers collocated on the beam. Then, a series of temporal spectrums are computed to demonstrate the evolution of EM admittance through fast Fourier transform of the mode-converted Lamb wave signals which are consecutively truncated in the time domain. When truncation time is relatively small, the corresponding temporal spectrum is governed by the characteristics of the input driving frequency. As truncation time becomes large, however, the modal characteristics of the notched beam play a crucial role in the temporal spectrum within the input driving frequency band. This implies that mode-converted Lamb waves reverberating on the beam contributes to the resonance of the beam. The root mean square values are computed for the temporal spectrums in the vicinity of each resonance frequency. The root mean square values increase monotonically with respect to truncation time for any resonance frequencies. Finally the implications of the numerical observation are discussed in the context of damage detection of a beam.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.35 no.5
• /
• pp.277-286
• /
• 2022

The elastic and inelastic responses obtained from the experimental and analytical results of two RC building structures under the service level earthquake (SLE) and maximum considered earthquake (MCE) in Korea were used to weinvestigate the characteristics of the mechanisms resisting shear and torsional behavior in torsionally unbalanced structures. Equations representing the interactive effect of translational drift and torsional deformation on the shear force and torsional moment were proposed. Because there is no correlation in the behavior between elastic and inelastic forces and strains, the incremental shear forces and incremental torsional moments were analyzed in terms of their corresponding incremental drifts and incremental torsional deformations with respect to the yield, unloading, and reloading phases around the maximum edge-frame drift. In the elastic combination of the two dominant modes, the translational drift mainly contributes to the shear force, whereas the torsional deformation contributes significantly to the overall torsional moment. However, this phenomenon is mostly altered in the inelastic response such that the incremental translational drift contributes to both the incremental shear forces and incremental torsional moments. In addition, the given equation is used to account for all phenomena, such as the reduction in torsional eccentricity, degradation of torsional stiffness, and apparent energy generation in an inelastic response.

• Journal of the Korean Society of Radiology
• /
• v.18 no.3
• /
• pp.187-201
• /
• 2024

Osteoporosis is a major health issue globally, often remaining undetected until a fracture occurs. To facilitate early detection, deep learning (DL) models were developed to classify osteoporosis using abdominal computed tomography (CT) scans. This study was conducted using retrospectively collected data from 3,012 contrast-enhanced abdominal CT scans. The DL models developed in this study were constructed for using image data, demographic/clinical information, and multi-modality data, respectively. Patients were categorized into the normal, osteopenia, and osteoporosis groups based on their T-scores, obtained from dual-energy X-ray absorptiometry, into normal, osteopenia, and osteoporosis groups. The models showed high accuracy and effectiveness, with the combined data model performing the best, achieving an area under the receiver operating characteristic curve of 0.94 and an accuracy of 0.80. The image-based model also performed well, while the demographic data model had lower accuracy and effectiveness. In addition, the DL model was interpreted by gradient-weighted class activation mapping (Grad-CAM) to highlight clinically relevant features in the images, revealing the femoral neck as a common site for fractures. The study shows that DL can accurately identify osteoporosis stages from clinical data, indicating the potential of abdominal CT scans in early osteoporosis detection and reducing fracture risks with prompt treatment.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.24 no.2
• /
• pp.79-86
• /
• 2014

This paper describes an integrity evaluation method for emergency diesel generator(EDG) and rotor part of EDG. EDG is a very important equipment in the nuclear power plant(NPP). EDG supplies electricity to the safety-related equipments for the safety shut down of NPP in an emergency situation of earthquake. The safety of the rotor part of EDG is also important during seismic impact from earthquake. The finite element modelling of the EDG including rotor part was constructed. The modal analysis of EDG was firstly performed. The first natural frequency was calculated and revealed higher than the cutoff frequency of seismic spectrum. Then the stress analysis was done to compare with the allowable stress. The safety of the rotor part was investigated by the finite element analysis of the rotor and journal bearing interaction to find film thickness and critical speed. The seismic load was applied to rotor part in a manner that the load was a weighted static load. Analysis results showed that the maximum stress was within the range of allowable stress and the film thickness is larger than the permissible minimum thickness, and the critical speed was out of the operating speed. Hence, the structural and dynamic integrity of EDG could be confirmed by the numerical analysis method used in this paper. However, dynamic analysis of a rotating rotor and supporting bearing with the seismic impact needs to be investigated in a more rigorous method since the seismic load to the rotating part complicates the behavior of rotating system.