• Korean Journal of Computational Design and Engineering
• /
• v.19 no.4
• /
• pp.423-433
• /
• 2014

For the proper decision making and responsibility enhancement for an unexpected accident in large-scale facilities, it is important to train operators or first responders to minimize potential human errors and consequences resulted from them. Simulation technologies, including human-computer interaction and virtual reality, enables personnel to participate in simulated hazardous situations with a safe, interactive, repetitive way to perform these training activities. For the development of accident response training simulator, it is necessary to define components comprising the simulator and to integrate them for the given training purpose. In this paper, we analyze requirements of the training simulator, derive key components, and design the training simulator. Based on the design, we developed a prototype training simulator and verified the simulator through experiments.

• Journal of Information Processing Systems
• /
• v.14 no.6
• /
• pp.1464-1479
• /
• 2018

In face recognition, sometimes the number of available training samples for single category is insufficient. Therefore, the performances of models trained by convolutional neural network are not ideal. The small sample face recognition algorithm based on novel Siamese network is proposed in this paper, which doesn't need rich samples for training. The algorithm designs and realizes a new Siamese network model, SiameseFacel, which uses pairs of face images as inputs and maps them to target space so that the $L_2$ norm distance in target space can represent the semantic distance in input space. The mapping is represented by the neural network in supervised learning. Moreover, a more lightweight Siamese network model, SiameseFace2, is designed to reduce the network parameters without losing accuracy. We also present a new method to generate training data and expand the number of training samples for single category in AR and labeled faces in the wild (LFW) datasets, which improves the recognition accuracy of the models. Four loss functions are adopted to carry out experiments on AR and LFW datasets. The results show that the contrastive loss function combined with new Siamese network model in this paper can effectively improve the accuracy of face recognition.

• Therapeutic Science for Rehabilitation
• /
• v.1 no.1
• /
• pp.57-70
• /
• 2012

Objective : This study was to evaluate the effects of a Neurofeedback(NFB) and Computer Assisted Cognitive Rehabilitation(CACR) training to improve on cognition and affected arm function in stroke subjects. Methods : Participants were randomly allocated to three groups: NFB(n=14), CACR(n=14) and control(n=16). All groups received occupational therapy and physical therapy for 5 session 30 minutes per week during 6 weeks. Also NFB and CACR group practiced additional NeuroComp training and RehaCom training for 30 sessions 30 minutes during 6 weeks. Results were evaluated by cognition, affected arm function. Results : There were significantly increased by CACR training that outcomes of MMSE-K(p<.05). And there weren't significantly difference by NFB and CACR training that outcomes of the affected arm function. And a difference between three groups wasn't found. Conclusion : The NFB and CACR training improves cognitive function. These results suggest that NFB and CACR training is feasible and suitable for individuals with stroke.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.323-326
• /
• 2019

2015 개정 교육과정이 고시됨에 따라 정규 교육과정을 중심으로 코딩 교육의 중요성이 강조되고 있다. 이에 따라 코딩 교육 강사가 요구되고 있는 시점이며, 이러한 강사는 여성새로일하기센터를 비롯한 민간 교육기관에서 코딩교육 강사 양성 프로그램이 운영되고 있다. 그러나 이러한 코딩교육 강사 양성 프로그램의 교육현황 및 교육 내용에 대한 연구는 이루어지고 있지 않은 실정이다. 따라서 본 논문에서는 여성가족부의 여성새로일하기센터에서 진행되고 있는 코딩 교육 프로그램의 사례를 분석하여 보고, 그 시사점을 찾고자 한다.

• International Journal of Computer Science & Network Security
• /
• v.21 no.6
• /
• pp.200-206
• /
• 2021

Fraud in e-commerce transaction increased in the last decade especially with the increasing number of online stores and the lockdown that forced more people to pay for services and groceries online using their credit card. Several machine learning methods were proposed to detect fraudulent transaction. Neural networks showed promising results, but it has some few drawbacks that can be overcome using optimization methods. There are two categories of learning optimization methods, first-order methods which utilizes gradient information to construct the next training iteration whereas, and second-order methods which derivatives use Hessian to calculate the iteration based on the optimization trajectory. There also some training refinements procedures that aims to potentially enhance the original accuracy while possibly reduce the model size. This paper investigate the performance of several NN models in detecting fraud in e-commerce transaction. The backpropagation model which is classified as first learning algorithm achieved the best accuracy 96% among all the models.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.7
• /
• pp.141-152
• /
• 2017

As the importance of technology increases, so too does its use in various areas. In education, technology is often used. However, due to teachers' lack of knowledge about technology, they often remain at the level of simple utilization, without applying it to learning. Thus, there is a growing need for Technological Pedagogical Content Knowledge (TPACK), which enables teachers to have knowledge about technology and use it appropriately given the content. Although TPACK studies are underway in many subjects, they suffer from the limited functionality of the included technology. To solve this problem, in this study, the range of technology in TPACK was extended to programming, and a TPACK-P model was developed to teach this expanded TPACK to pre-service teachers. To verify the effectiveness of this model, the TPACK-P training program developed during the 15th classes was applied to 19 pre-service teachers. We used Park and Kang (2014) as a tool to measure these teachers' TPACK before and after treatment to observe any changes. The results showed that the TPACK-P education program showed statistically significant improvement in all areas except Pedagogical Content Knowledge(PCK). Compared with the ICT-based TPACK training program, which was administered to a control group, the TPACK-P training program proved to be more effective in the development of Technological Pedagogical Knowledge(TPK) and TPACK among pre-service teachers.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3668-3684
• /
• 2021

Video action recognition is widely used in video surveillance, behavior detection, human-computer interaction, medically assisted diagnosis and motion analysis. However, video action recognition can be disturbed by many factors, such as background, illumination and so on. Two-stream convolutional neural network uses the video spatial and temporal models to train separately, and performs fusion at the output end. The multi segment Two-Stream convolutional neural network model trains temporal and spatial information from the video to extract their feature and fuse them, then determine the category of video action. Google Xception model and the transfer learning is adopted in this paper, and the Xception model which trained on ImageNet is used as the initial weight. It greatly overcomes the problem of model underfitting caused by insufficient video behavior dataset, and it can effectively reduce the influence of various factors in the video. This way also greatly improves the accuracy and reduces the training time. What's more, to make up for the shortage of dataset, the kinetics400 dataset was used for pre-training, which greatly improved the accuracy of the model. In this applied research, through continuous efforts, the expected goal is basically achieved, and according to the study and research, the design of the original dual-flow model is improved.

• The Journal of Korea Robotics Society
• /
• v.18 no.3
• /
• pp.346-351
• /
• 2023

With the release of numerous open driving datasets, the demand for domain adaptation in perception tasks has increased, particularly when transferring knowledge from rich datasets to novel domains. However, it is difficult to solve the change 1) in the sensor domain caused by heterogeneous LiDAR sensors and 2) in the environmental domain caused by different environmental factors. We overcome domain differences in the semi-supervised setting with 3-stage model parameter training. First, we pre-train the model with the source dataset with object scaling based on statistics of the object size. Then we fine-tine the partially frozen model weights with copy-and-paste augmentation. The 3D points in the box labels are copied from one scene and pasted to the other scenes. Finally, we use the knowledge distillation method to update the student network with a moving average from the teacher network along with a self-training method with pseudo labels. Test-Time Augmentation with varying z values is employed to predict the final results. Our method achieved 3rd place in ECCV 2022 workshop on the 3D Perception for Autonomous Driving challenge.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.4 no.1
• /
• pp.35-42
• /
• 1999

This paper describes the development of an interactive computer graphic software for the education and training of the power system fault analysis. The developed software is designed to increase the understanding of the fault analysis with ease and it is composed of the windows, graphic icons, and graphic representations for user-friendly environments. Specially an interactive scheme is given for user to simulate the fault analysis under the variety conditions. With this function, user can acquire the basic concepts of the power system fault study as well as the understanding of the impacts on the system by some faults. The proposed software is tested on a 16-bus sample system. The software will be useful for the education and training and training of the power system fault analysis.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.3
• /
• pp.71-77
• /
• 2018

The purpose of this study was to investigate the effectiveness of elastic band resistance training for muscle strength among community-dwelling older adults. The systematic review and meta-analysis was conducted by following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA). Data were pooled using fixed effect models. Sit to stand, arm curl, and grip strength were analyzed for main effects. Heterogeneity between studies was assessed using the I2 statistics and publication bias was evaluated by funnel plots. Twelves studies were included representing 611 participants. Elastic band resistance training was effective for lower (d=3.89, 95% CI: 3.03, 4.75) and upper extremity muscle strength (d=4.08, 95% CI: 2.94, 5.23). Heterogeneity was moderate and no significant publication bias was detected. Based on these findings, there is clear evidence that elastic band resistance training has significant positive effects on muscle strength among community-dwelling older adults. Further study will be needed to perform subgroup analysis using number of sessions and exercise intensity as predictors.