• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.102-105
• /
• 2019

In the classical single-image super-resolution (SISR) reconstruction method using convolutional neural networks, the extracted features are not fully utilized, and the training time is too long. Aiming at the above problems, we proposed an improved SISR method based on a residual network. Our proposed method uses a feature fusion technology based on improved residual blocks. The advantage of this method is the ability to fully and effectively utilize the features extracted from the shallow layers. In addition, we can see that the feature fusion can adaptively preserve the information from current and previous residual blocks and stabilize the training for deeper network. And we use the global residual learning to make network training easier. The experimental results show that the proposed method gets better performance than classic reconstruction methods.

• Asia Marketing Journal
• /
• v.19 no.3
• /
• pp.75-98
• /
• 2017

Our study is trying to investigate the perceived training effectiveness by applying the theory of planned behavior (TPB) and Technological Acceptance Model (TAM) and intend to examine the effects of individual differences on perceived training effectiveness and performance of individuals. The main purpose is to evaluate the perceived training effectiveness, and role of individual differences in terms of learning. The results of this study supported all the hypothesis that participants with higher level of creative self-efficacy, intrinsic motivation, creativity and emotional intelligence (EI) will have greater inclinations to learn. Results showed that perceive training effectiveness is positively related to training transfer and training transfer increase the performance of individuals. Study results significantly agree with the theory of planned behavior (TPB) which was applied to measure the perceived training effectiveness and suggest trainee's perception of usefulness, ease and benefits enhance learning dimensions of participants that make any program effective. The study has highlighted a number of issues that influence the perceived training effectiveness.

• International Journal of Computer Science & Network Security
• /
• v.22 no.2
• /
• pp.214-222
• /
• 2022

Convolutional Neural networks (CNNs) are a category of deep learning networks that have proven very effective in computer vision tasks such as image classification. Notwithstanding, not much has been seen in its use for remote sensing image classification in developing countries. This is majorly due to the scarcity of training data. Recently, transfer learning technique has successfully been used to develop state-of-the art models for remote sensing (RS) image classification tasks using training and testing data from well-known RS data repositories. However, the ability of such model to classify RS test data from a different dataset has not been sufficiently investigated. In this paper, we propose a deep CNN model that can classify RS test data from a dataset different from the training dataset. To achieve our objective, we first, re-trained a ResNet-50 model using EuroSAT, a large-scale RS dataset to develop a base model then we integrated Augmentation and Ensemble learning to improve its generalization ability. We further experimented on the ability of this model to classify a novel dataset (Nig_Images). The final classification results shows that our model achieves a 96% and 80% accuracy on EuroSAT and Nig_Images test data respectively. Adequate knowledge and usage of this framework is expected to encourage research and the usage of deep CNNs for land cover mapping in cases of lack of training data as obtainable in developing countries.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.209-216
• /
• 2019

Over the past decade, researchers were able to solve complex medical problems as well as acquire deeper understanding of entire issue due to the availability of machine learning techniques, particularly predictive algorithms and automatic recognition of patterns in medical imaging. In this study, a technique called transfer learning has been utilized to classify Magnetic Resonance (MR) images by a pre-trained Convolutional Neural Network (CNN). Rather than training an entire model from scratch, transfer learning approach uses the CNN model by fine-tuning them, to classify MR images into Alzheimer's disease (AD), mild cognitive impairment (MCI) and normal control (NC). The performance of this method has been evaluated over Alzheimer's Disease Neuroimaging (ADNI) dataset by changing the learning rate of the model. Moreover, in this study, in order to demonstrate the transfer learning approach we utilize different pre-trained deep learning models such as GoogLeNet, VGG-16, AlexNet and ResNet-18, and compare their efficiency to classify AD. The overall classification accuracy resulted by GoogLeNet for training and testing was 99.84% and 98.25% respectively, which was exceptionally more than other models training and testing accuracies.

• The Journal of Korean Association of Computer Education
• /
• v.5 no.4
• /
• pp.147-154
• /
• 2002

In this paper, we investigated the effects of response technique which uses free tokens in the computer training for the mentally retarded person. We set wrong line entering. wrong space number entering, wrong line-type designation and wrong cell merger as the target behaviors in the research. The experiment period of the research was 36 sessions altogether, and the period was divided into four steps such as basic period I. treatment period I, basic period II, and treatment period II. We also conducted a follow up experiment to identify whether the experiment's effect is still maintained two weeks after the experiment finished. We know that the response cost technique which uses free tokens to the mentally retarded person is efficient in the computer training.

• International journal of advanced smart convergence
• /
• v.9 no.1
• /
• pp.202-208
• /
• 2020

Body balance necessary for ordinary daily activities can be undermined by diverse causes. In this study, as a way to control such a problem, we have produced smart insole as a wearable device in the form of insole and developed analysis software evaluating body balance, which measures ground reaction force applied to each area of sole and Center of Pressure (COP). The software visualized changes in COP positions while a user was moving and average COP positions, and it is also capable of measuring the COP values in the Anterior-Posterior (AP) and Medial-Lateral (ML) areas of feet. Through gait analysis, it can analyze the time of walking, strides, speed, COP trajectory while walking, etc. In addition, we have developed training contents for body balance improvement designed in consideration of Y-Balance Test and Timed Up and Go (TUG) Test. They were established in virtual reality similar to daily living environment so that people can expect more effective training results regardless of places.

• Journal of The Korean Association of Information Education
• /
• v.13 no.1
• /
• pp.23-30
• /
• 2009

In this study, we proposed a pre-service teacher training program that assists efficiently the conceptual comprehension and teaching skill development of computer science education. The program is integrated with Protype Theory and Example Theory based on cognitive psychology. And also the real teaching activities based on Unplugged Learning are provided for conceptual comprehension of computer science education as well as for learning computer science. This program was applied to 31 pre-service teachers for one semester. The results show that this program is very effective for pre-service teacher training of computer science education.

• Progress in Superconductivity and Cryogenics
• /
• v.24 no.3
• /
• pp.24-29
• /
• 2022

One of the recent proposals with nuclear magnetic resonance (NMR) is a multi-bore NMR which consists of array of magnets which could present possibilities to quickly cope with pandemic virus by multiple inspection of virus samples. Low temperature superconductor (LTS) can be a candidate for mass production of the magnet due to its low price in fabrication as well as operation by applying the helium zero boil-off technology. However, training feature of LTS magnet still hinders the low cost operation due to multiple boil-offs during premature quenches. Thus in this paper, LTS magnet with low mechanical stress is designed targeting the "training-free" LTS magnet for mass production of magnet array for multi-bore NMR. A thorough process of an LTS magnet design is conducted, including the analyses as the followings: electromagnetics, mechanical stress, cryogenics, stability, and protection. The magnet specification was set to 4.7 T in a winding bore of 85 mm, corresponding to the MR frequency of 200 MHz. The stress level is tolerable with respect to the wire yield strength and epoxy crack where mechanical disturbance is less than the minimum quench energy.

• Journal of IKEEE
• /
• v.22 no.3
• /
• pp.816-821
• /
• 2018

Recently, machine learning algorithms based on artificial neural networks started to be used widely as classifiers in the field of EEG research for emotion analysis and disease diagnosis. When a machine learning model is used to classify EEG data, if training data is composed of only data having similar characteristics, classification performance may be deteriorated when applied to data of another group. In this paper, we propose a method to construct training data set by selecting several groups of data using semi-supervised learning algorithm to improve these problems. We then compared the performance of the two models by training the model with a training data set consisting of data with similar characteristics to the training data set constructed using the proposed method.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.9
• /
• pp.107-112
• /
• 2016

In this paper, we propose the effects of military cardiopulmonary resuscitation(CPR) on the quality of debriefing and feedback device training. The key idea of combination debriefing and feedback device training is to maximize effects of CPR. The participants of the research were non-medic soldiers in ROK army, and had not undergone any professional CPR training before. Each group of soldier was randomized to perform of military CPR by using training method in each group. After 5 minutes of performing CPR, each D, F, DF group showed significant improvement in CPR performance. When comparing each group, the rate of success in CPR performance in DF group was significantly higher than that of F group with the average difference of 11.160(p<.01) points. In summation, the training programs that DF received seemed to be more efficient and effective than that of D and F. The fatigue level was evaluated by comparing the lactate concentration in blood after performing CPR. Through this experiment, we show that the training programs that DF received is more efficient and effective than that of D and F.