• 한국산학기술학회논문지
• /
• 제18권4호
• /
• pp.341-348
• /
• 2017

본 연구에서는 척추측만증 환자의 척추 만곡 위치와 방향이 자세 균형에 미치는 영향을 평가하였다. 총 15명의 실험대상자를 흉추 만곡 그룹, 요추 만곡 그룹, 이중 만곡 그룹으로 분류하여 연구를 진행하였다. 초음파 기반 동작 분석 시스템과 압력 분포 시스템을 사용하여 환자의 동적 체간 움직임(요추, 흉요추, 하흉추, 상흉추에서의 각도 변화)과 족저 압력 분포(최대힘, 최대압력)을 측정하였다. 측정 결과를 통해, 특발성 척추측만증 환자의 동적 체간 움직임과 족저 압력 분포 모두 척추 만곡의 발생 부위와 방향에 따라 비대칭적으로 각도와 압력이 증가하면서 자세 불균형이 발생하는 것을 알 수 있었다. 또한, 단일 만곡과 이중 만곡을 가진 그룹 간의 자세 균형 패턴에서의 차이를 확인할 수 있었다. 추후 연구에서는 본 연구에서의 결과를 기반으로 척추측만증 환자의 자세 조절 능력과 체간 균형을 향상시키고 척추측만을 치료하는데 도움을 줄 수 있는 재활 훈련 장치를 개발하고자 한다.

• 한국음향학회지
• /
• 제40권5호
• /
• pp.466-472
• /
• 2021

화자 전환 검출은 대화 중에 발성 화자가 다른 사람으로 바뀌는 시점을 검출하는 것을 의미한다. 이 과정에서 화자 중복, 화자 정보 표기의 부정확성, 데이터 불균형 등으로 화자가 바뀌는 순간을 검출하는 데 어려움이 발생한다. 본 논문에서는 이러한 문제를 해결하기 위해 음성 인식에 널리 사용되는 TIMIT 데이터를 가공하여 충분한 양의 훈련 데이터를 얻었으며, 화자가 겹치는지를 파악한 후에 화자 전환 여부를 판단하였다. 본 논문에서는 화자 겹침을 고려한 화자 전환 검출 시스템을 구축하기 위하여 다양한 접근법을 사용하여 성능을 평가하고 검증했다. 그 결과 화자 겹칩 영역을 제거하기 위해 X-Vector 구조와 유사한 형태의 검출 시스템과 화자 전환 검출 시스템을 모델링하기 위한 Bi-LSTM 모델을 제안하였다. 실험 결과 기준 시스템보다 상대적으로 각각 4.6 %, 13.8 % 성능 향상을 확인하였다. 또한, 실험 결과를 기반으로 텍스트 정보와 화자 정보 등을 고려한다면 좀 더 강인한 화자 전환 검출 시스템을 구축할 수 있을 것으로 판단한다.

• 한국융합학회논문지
• /
• 제10권5호
• /
• pp.35-42
• /
• 2019

최근 사물인터넷과 다양한 웨어러블 기기들이 등장하면서 인터넷 기술은 보다 편리하게 정보를 얻고 업무를 수행하는데 기여하고 있으나 인터넷이 다양한 부분에 이용되면서 공격에 노출되는 Attack Surface 지점이 증가하고 있으며 개인정보 획득, 위조, 사이버 테러 등 부당한 이익을 취하기 위한 목적의 네트워크 침입 시도 또한 증가하고 있다. 본 논문에서는 네트워크에서 발생하는 트래픽에서 비정상적인 행동을 분류하기 위한 희소클래스의 분류 성능을 개선하는 특징선택을 제안한다. UNSW-NB15 데이터셋은 다른 클래스에 비해 상대적으로 적은 인스턴스를 가지는 희소클래스 불균형 문제가 발생하며 이를 제거하기 위해 언더샘플링 방법을 사용한다. 학습 알고리즘으로 SVM, k-NN 및 decision tree를 사용하고 훈련과 검증을 통하여 탐지 정확도와 RMSE가 우수한 조합의 서브셋들을 추출한다. 서브셋들은 래퍼 기반의 실험을 통해 재현률 98%이상의 유효성을 입증하였으며 DT_PSO 방법이 가장 우수한 성능을 보였다.

• PNF and Movement
• /
• 제12권4호
• /
• pp.243-248
• /
• 2014

Purpose: The transverse abdominis and themultifidus muscle are located in the core. They surround one's trunk and help in body stabilization. Specifically, they control spine articulation to maintain posture and balance. Therefore, weakened deep muscle in the trunk may cause spinal malalignment. This study aims to compare the correlation between the thickness of the transverse abdominis and the multifidus muscle and the spine alignment among college students in their 20s. Methods: This study measured the thickness of the transverse abdominis and the multifidus muscle of 42 healthy college students in their 20s using ultrasonic waves. The thickness of the muscle was measured for the length of the cross-section except for fascia. The thickness of the left and right muscles was measured, and the mean value was calculated. As the thickness of the transverse abdominis can increase because of pressure during exhalation, it was measured at the last moment of exhalation. Spinal alignment was measured by the kyphosis angle, lordosis angle, pelvic tilt, trunk inclination, lateral deviation, trunk imbalance, and surface rotation using Formetric III, which is a three-dimensional imaging equipment. They were measured for three times, and the mean values were calculated. The general characteristics of the subjects were analyzed using descriptive statistics. The correlations between each factor were analyzed using Pearson's correlation analysis. Results: The transverse abdominis showed asignificant correlation with trunk inclination (p<.05). The multifidus muscle showed a significant positive correlation with pelvic tilt and a negative correlation with surface rotation (p<.05). Conclusion: The thickness of transverse abdominis and the multifidus muscle appears to influence spinal alignment. Specifically, the multifidus muscle, which plays an important role on the sagittal plane, influences surface rotation, thus making it an important muscle for scoliosis patients. Therefore, a strengthening training program for the transverse abdominis and the multifidus muscle is necessary according to specific purposes among adults with spinal malalignment.

• East Asian Economic Review
• /
• 제26권1호
• /
• pp.49-91
• /
• 2022

Recent years have witnessed renewed interest in Africa and public diplomacy has emerged as the vital tool being used to cultivate these relations. China has been leading in pursuing stronger economic partnership with Africa while middle powers such as Korea are also intensifying engagement with the continent. While previous studies have analyzed the implications of China's activities in Africa on advanced powers, none has examined them from the paradigm of middle powers. This study fills this gap by assessing China's activities in Africa, their economic engagement and implications for Korea's interest in Africa. The analysis is qualitative based on secondary data from various sources and literature. The study shows that China's public diplomacy strategy involves a high degree of innovation and has evolved to encompass new tools and audiences. China has institutionalized a cooperative model that permeates many aspects of governance institutions in Africa, enabling it to strengthen their relations. This could also be helping China to adjust faster leadership transitions in Africa. Whereas the US is still the most influential country in Africa, China is influential in economic policies and has outstripped the US in infrastructure diplomacy. This could be because African policy makers align more with China's economic model than the US' mainstream economics. Chinese aid to Africa has been diversified to social sectors that are more responsive to the needs of Africa. Trade and investment relations between China and Africa have deepened, but so does trade imbalance since 2010. China mainly imports natural resources and raw materials from Africa. But this product portfolio is not different from Korea and the US. China's energetic insertion in Africa using various strategies has significant implications for countries with ambitions in Africa. Korea can achieve its ambitions in Africa by focusing resources in areas it can leverage its core strengths-such as education and vocational training, environmental policy and development cooperation.

• Smart Structures and Systems
• /
• 제30권6호
• /
• pp.613-626
• /
• 2022

Guaranteeing the quality and integrity of structural health monitoring (SHM) data is very important for an effective assessment of structural condition. However, sensory system may malfunction due to sensor fault or harsh operational environment, resulting in multiple types of data anomaly existing in the measured data. Efficiently and automatically identifying anomalies from the vast amounts of measured data is significant for assessing the structural conditions and early warning for structural failure in SHM. The major challenges of current automated data anomaly detection methods are the imbalance of dataset categories. In terms of the feature of actual anomalous data, this paper proposes a data anomaly detection method based on data-level and deep learning technique for SHM of civil engineering structures. The proposed method consists of a data balancing phase to prepare a comprehensive training dataset based on data-level technique, and an anomaly detection phase based on a sophisticatedly designed network. The advanced densely connected convolutional network (DenseNet) and Transformer encoder are embedded in the specific network to facilitate extraction of both detail and global features of response data, and to establish the mapping between the highest level of abstractive features and data anomaly class. Numerical studies on a steel frame model are conducted to evaluate the performance and noise immunity of using the proposed network for data anomaly detection. The applicability of the proposed method for data anomaly classification is validated with the measured data of a practical supertall structure. The proposed method presents a remarkable performance on data anomaly detection, which reaches a 95.7% overall accuracy with practical engineering structural monitoring data, which demonstrates the effectiveness of data balancing and the robust classification capability of the proposed network.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제12권8호
• /
• pp.235-242
• /
• 2023

최근 들어 영유아를 대상으로 한 비대면 상담이 증가함에 따라 감정인식 보조 도구로 CNN기반 딥러닝 모델을 많이 사용하고 있다. 하지만 대부분의 감정인식 모델은 성인 데이터 위주로 학습되어 있어 영유아 및 청소년을 대상으로 적용하기에는 성능상의 제약이 있다. 본 논문에서는 이러한 성능제약의 원인을 분석하기 위하여 XAI 기법 중 하나인 LIME 기법을 통해 성인 대비 영유아와 청소년의 감정인식을 위한 얼굴 표정의 특징을 분석한다. 뿐만 아니라 남녀 집단에도 동일한 실험을 수행함으로써 성별 간 얼굴 표정의 특징을 분석한다. 그 결과로 연령대별 실험 결과와 성별별 실험 결과를 CNN 모델의 사전 훈련 데이터셋의 데이터 분포를 바탕으로 설명하고 균형 있는 학습 데이터의 중요성을 강조한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권6호
• /
• pp.1530-1544
• /
• 2023

Data monitoring is an important foundation of modern science. In most cases, the monitoring data is time-series data, which has high application value. The deep learning algorithm has a strong nonlinear fitting capability, which enables the recognition of time series by capturing anomalous information in time series. At present, the research of time series recognition based on deep learning is especially important for data monitoring. Deep learning algorithms require a large amount of data for training. However, abnormal sample is a small sample in time series, which means the number of abnormal time series can seriously affect the accuracy of recognition algorithm because of class imbalance. In order to increase the number of abnormal sample, a data augmentation method called GANBATS (GAN-based Bi-LSTM and Attention for Time Series) is proposed. In GANBATS, Bi-LSTM is introduced to extract the timing features and then transfer features to the generator network of GANBATS.GANBATS also modifies the discriminator network by adding an attention mechanism to achieve global attention for time series. At the end of discriminator, GANBATS is adding averagepooling layer, which merges temporal features to boost the operational efficiency. In this paper, four time series datasets and five data augmentation algorithms are used for comparison experiments. The generated data are measured by PRD(Percent Root Mean Square Difference) and DTW(Dynamic Time Warping). The experimental results show that GANBATS reduces up to 26.22 in PRD metric and 9.45 in DTW metric. In addition, this paper uses different algorithms to reconstruct the datasets and compare them by classification accuracy. The classification accuracy is improved by 6.44%-12.96% on four time series datasets.

• 방사선산업학회지
• /
• 제18권2호
• /
• pp.117-126
• /
• 2024

This study aims to identify high exposure tasks among the tasks performed in domestic nuclear power plants as a basis for developing training programs to improve the efficiency of workers' work. To this end, we first analyzed the exposure status of radiation work in domestic nuclear power plants. Radiation tasks in nuclear power plants were categorized, collective doses were investigated, and the collective doses were calculated based on the collective doses, and representative high exposure tasks were identified. We found that the collective and individual doses in domestic nuclear power plants are continuously decreasing, but there is an imbalance of exposure among workers. In terms of work classification, nuclear power plants are managed in 236 work codes based on light water reactors and 181 work codes based on heavy water reactors, depending on the work equipment and location. Among the total work codes, 23 codes have an annual average dose exceeding 10 μSv, and based on this, 10 representative high exposure tasks were derived. The representative high exposure tasks were selected as S/G nozzle dam work, S/G debris removal work, nuclear instrumentation system, S/G eddy current detection work, and insulation work. The results of this study are expected to serve as an important basis for reducing the exposure of workers in nuclear power plants and improving work efficiency.

• 원예과학기술지
• /
• 제33권2호
• /
• pp.155-165
• /
• 2015

우량 측지묘(측지수 : 10개 이상)를 $3.0{\times}1.0m$ 거리로 심고 수고를 2.5m 정도로 한 세장방추형과 수고를 3.5m로 높인 키큰방추형으로 키우면서 8년간 수체생장, 생산성 및 과실품질을 비교하였다. 재식 4년차에 키큰방추형의 수관용적이 세장방추형보다 커지기 시작하여 5년차에는 주당 수관용적이 키큰방추형은 세장방추형에 비해 25% 정도 더 컸다. 8년 동안의 10a당 누적수량은 키큰방추형이 세장방추형에 비해 14% 정도 증수되었다. 더군다나 재식 5년차 이후로 두 시험구 모두 해거리 및 갈색무늬병이 발생하여, 수세가 종종 불안정하였는데, 키큰방추형의 생산량 감소 및 수세 불안정 정도는 세장방추형보다 덜하였다. 가용성고형물 함량과 착색은 재식 5년차에 키큰방추형이 세장방추형보다 증가하였는데, 이는 키큰방추형의 측지 및 착과가 세장방추형보다 균일하게 배치되면서 수관 내 광투과율이 증가되었기 때문으로 생각되었다. 결론적으로 국내에서 '후지'/M.9 사과나무를 333주/10a 이상으로 재식할 경우 세장방추형의 수고를 3.5m로 높이는 것이 광투과율, 생산량 및 과실품질 측면에서 관행의 세장방추형보다 나을 것으로 판단되었다.