• 한국인터넷방송통신학회논문지
• /
• 제21권6호
• /
• pp.111-116
• /
• 2021

초분광 영상은 적외선 영역의 전자기파 대역을 수백 개의 파장으로 나누어 영상화한 데이터로 다양한 분야에서 물체를 찾거나 분류하는 것에 활용된다. 최근에는 딥러닝을 사용하여 분류하는 방법이 주목받고 있지만 초분광 영상 데이터의 특성으로 인해 초분광 영상을 학습 데이터로 사용하기 위해서는 기존의 가시광 영상과는 다른 처리 기법이 필요하다. 이를 위해 초분광 큐브에서 특정 파장의 영상을 선택하여 Ground Truth 작업을 수행하고 환경정보를 포함하여 데이터를 관리하는 소프트웨어를 개발하였다. 본 논문에서는 해당 소프트웨어의 구성과 기능에 대하여 설명한다.

• ETRI Journal
• /
• 제45권5호
• /
• pp.795-810
• /
• 2023

Unmanned aerial vehicles (UAV) and ground vehicles (UGV) require advanced video analytics for various tasks, such as moving object detection and segmentation; this has led to increasing demands for these methods. We propose a zero-shot video object segmentation method specifically designed for UAV and UGV applications that focuses on the discovery of moving objects in challenging scenarios. This method employs a background memory model that enables training from sparse annotations along the time axis, utilizing temporal modeling of the background to detect moving objects effectively. The proposed method addresses the limitations of the existing state-of-the-art methods for detecting salient objects within images, regardless of their movements. In particular, our method achieved mean J and F values of 82.7 and 81.2 on the DAVIS'16, respectively. We also conducted extensive ablation studies that highlighted the contributions of various input compositions and combinations of datasets used for training. In future developments, we will integrate the proposed method with additional systems, such as tracking and obstacle avoidance functionalities.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2013년도 추계학술대회
• /
• pp.1021-1023
• /
• 2013

현재 항공기, 자동차 훈련용 시뮬레이터는 지상에 설치된 시뮬레이터에 가상의 상황을 구현하여 교육생에게 다양한 교육훈련 내용을 제공하고 교관은 다른 공간에서 교육훈련에 대한 내용을 모니터링하고 교육훈련에 필요한 내용은 바로 지시하여 교육훈련의 효과를 극대화하고 있다. 하지만 교육생이 실제 항공기나 자동차를 탑승하여 훈련을 진행 할 경우 지상의 교관은 항공기나, 자동차에 대한 내용을 지상에서 모니터링 할 수 없고 교육 종료 후 교육훈련 내용에 대한 평가가 쉽지 않아 교육생에게 양질의 교육이 어렵다. 본 논문에서는 GPS와 9-axis sensor를 탑재한 항공기, 차량에 GPS정보와 sensor 정보를 실시간으로 수집하여 항공기와 차량의 상태를 3D 시뮬레이션 영상을 구현하여 중앙의 관제센터에서 모니터링하여 교육훈련에 활용하고 GPS 정보와 sensor 정보를 파일로 저장하여 교육훈련 종료 후 저장된 GPS정보와 sensor 정보를 다시 3D 영상을 구현하여 교육생과 교관이 교육훈련 내용에 대하여 평가와 분석을 할 수 있는 시뮬레이션 소프트웨어를 개발하고자 한다.

• The Journal of Korean Physical Therapy
• /
• 제27권2호
• /
• pp.124-127
• /
• 2015

Purpose: The emphasis on gait rehabilitation after stroke depends on training support through the lower limbs, balance of body mass over the changing base of support. However, muscle weakness, lack of control of lower limb, and poor balance can interfere with training after stroke. For this case study report, a wearable robot orthosis was applied to stroke patients in order to verify its actual applicability on balance and gait ability in the clinical field. Methods: Two stroke patients participated in the training using the wearable robot orthosis. Wearable robot orthosis provides patient-initiated active assistance contraction during training. Training includes weight shift training, standing up and sitting down, ground walking, and stair up and down Training was applied a total of 20 times, five times a week for 4 weeks, for 30 minutes a day. Gait ability was determined by Stance phase symmetry profile, Swing phase symmetry profile, and velocity using the GAITRite system. Balance ability was measured using the Biodex balance system. Results: Subjects 1, 2 showed improved gait and balance ability with mean individual improvement of 72.4% for velocity, 19.4% for stance phase symmetry profile, 9.6% for swing phase symmetry profile, and 13.6% for balance ability. Conclusion: Training utilizing a wearable robot orthosis can be useful for improvement of the gait and balance ability of stroke patients.

• 식물병연구
• /
• 제30권1호
• /
• pp.34-49
• /
• 2024

The development of sustainable smut management techniques requires an understanding of the impacts of smut on sugarcane growth and the relationships between smut intensity and meteorological variables, varieties, and crop types. Thus, assessments were made with the objectives to 1) determine the effect of smut on the above-ground growth of sugarcane, and 2) quantify the association of smut with weather variables, varieties and crop types. The effect of smut on above-ground growth was assessed in six fields planted with NCo 334 (wider coverage) having 6 months of age in Fincha and Metehara fields in 2021. Data on above-ground growth were taken from 20 randomly selected smut-affected and healthy stools from each field. Besides, 6 years' data (2015 to 2021) on the numbers of smut-affected stools and smut whips of 79 fields were collected. Furthermore, 10 years' (2011 to 2021) weather data were acquired from the sugar plantations. The results demonstrated reduction in the above-ground growth of sugarcane in the range of 18.39% and 73.42% due to smut. In addition, weather variables explained about 68.48% and 66.58% of the variability in the number of smut-affected stools and whips respectively. Smut intensity increased with crop types for susceptible varieties. The tight association between the smut epidemic and crop types, varieties, and weather, implied that these parameters must be carefully considered in management decisions. Continuous monitoring of smut disease, meteorological variables, varieties, and crop types in all the sugarcane plantations could be done as a part of integrated smut management in the future.

• 한국공간구조학회논문집
• /
• 제20권2호
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• The Journal of Korean Physical Therapy
• /
• 제31권5호
• /
• pp.292-297
• /
• 2019

Purpose: This study examined the effects of backward walking training with task orientation on the functional walking ability of children with cerebral palsy. Methods: This study was a single-blinded, randomized controlled trial with a crossover design conducted at a single rehabilitation facility with cross-over to the other intervention arm following a two-week break. For a total of 12 children with spastic hemiplegia cerebral palsy, the forward walking training group (n=6) underwent training three times a week for three weeks, 40 minutes a day, and the backward walking training group (n=6) was also trained under the same conditions. To identify the functional walking ability, variables, such as the walking speed, stride length, and step length, were measured using a walk analyzer (OptoGait, Microgate S.r.l, Italy). Results: Both groups showed significant increases in walking speed, stride length, and step length (p<0.01). The backward walking group showed more significant improvement in the walking speed from pre- to post-test (p<0.05). The gait characteristics were similar in the two groups (stride length and step length) but the walking speed in the backward walking group showed a mean difference between the positive effects higher than the forward walking group. Conclusion: Task-oriented backward walking training, which was conducted on the ground, may be a more effective treatment approach for improving the walking functions of spastic hemiplegia children than forward walk training.

• 대한물리치료과학회지
• /
• 제29권2호
• /
• pp.20-27
• /
• 2022

Background: The purpose of the present study was to determine whether balance training combined with light touch improves the postural sway and gait speed in stroke patients. Design: Randomized Controlled Trial Study Methods: The current study included 40 stroke patients, who were randomly divided into two groups: the balance training with light touch (LT) group (n=20) and the balance training with heavy touch (HT) group (n=20). Both groups performed balance training on a stable and unstable ground. Additionally, the LT group provided light touch stimulation, and the HT group provided heavy touch stimulation. All the participants underwent 20 sessions of STS training (thirty minutes, five days per week for four weeks). In this study, postural sway was measured using a force platform, and walking speed was measured using a 10 meter walk test. Results: The improvements in postural sway and gait speed were observed to be significantly greater in the LT group, compared to the HT group (p<0.05). Conclusion: The results of the current study imply that balance training combined with light touch is a beneficial and effective therapeutic modality that can be employed to improve the postural sway and gait speed in stroke patients.

• 대한원격탐사학회지
• /
• 제20권5호
• /
• pp.353-359
• /
• 2004

This study proposes an approach to unsupervisedly estimate the number of classes and the parameters of defining the classes in order to train the classifier. In the proposed method, the image is segmented using a spatial region growing based on hierarchical clustering, and fuzzy training is then employed to find the sample classes that well represent the ground truth. For cluster validation, this approach iteratively estimates the class-parameters in the fuzzy training for the sample classes and continuously computes the log-likelihood ratio of two consecutive class-numbers. The maximum ratio rule is applied to determine the optimal number of classes. The experimental results show that the new scheme proposed in this study could be used to select the regions with different characteristics existed on the scene of observed image as an alternative of field survey that is so expensive.

• 한국항공운항학회지
• /
• 제28권4호
• /
• pp.69-81
• /
• 2020

Since airline safety depends significantly on the role of pilot, the importance of airlines' training qualified pilots and investing in the training program grows. This study aims to analyze the determinants of the commercial airline pilot competencies during the transition course. To this purpose, one-year training evaluation results of 215 new airline pilots in the JET transition course of Company K in 2019 are studied with correlation coefficient analysis, factor analysis, ANOVA, and regression analysis. Undergraduate major and Uljin-trained pilots show higher final-stage scores in PROC, AUTO, and CRM. Due to Uljin-trained pilots' higher first-stage scores, their score increase during the course is limited. Uljin-trained pilots' MANUAL scores in the final stage are lower than those of undergraduate major and overseas-trained pilots. Influence of trained location is found greater than English competency, undergraduate GPA, and ground school scores on commercial airline pilot competency evaluation scores during the program. The results will be useful in establishing scenario-based training focusing on different background to apply competency-based training and assessment.