• Design & Manufacturing
• /
• v.15 no.3
• /
• pp.1-6
• /
• 2021

An artificial neural network model based on a deep learning algorithm is known to be more accurate than humans in image classification, but there is still a limit in the sense that there needs to be a lot of training data that can be called big data. Therefore, various techniques are being studied to build an artificial neural network model with high precision, even with small data. The transfer learning technique is assessed as an excellent alternative. As a result, the purpose of this study is to develop an artificial neural network system that can classify burr images of light guide plate products with 99% accuracy using transfer learning technique. Specifically, for the light guide plate product, 150 images of the normal product and the burr were taken at various angles, heights, positions, etc., respectively. Then, after the preprocessing of images such as thresholding and image augmentation, for a total of 3,300 images were generated. 2,970 images were separated for training, while the remaining 330 images were separated for model accuracy testing. For the transfer learning, a base model was developed using the NASNet-Large model that pre-trained 14 million ImageNet data. According to the final model accuracy test, the 99% accuracy in the image classification for training and test images was confirmed. Consequently, based on the results of this study, it is expected to help develop an integrated AI production management system by training not only the burr but also various defective images.

• Journal of Korea Multimedia Society
• /
• v.25 no.2
• /
• pp.237-246
• /
• 2022

Among the technologies of the 4th industrial revolution, drones that have grown rapidly and are being used in various industries can be operated by the pilot directly or can be operated automatically through programming. In order to be controlled by a pilot or to operate automatically, it is essential to predict and analyze the optimal path for the drone to move without obstacles. In this paper, after securing and analyzing the pilot training dataset through the unmanned aerial vehicle piloting training platform designed through prior research, the profile of the dataset that should be preceded to search and derive the optimal route of the unmanned aerial vehicle was designed. The drone pilot training data includes the speed, movement distance, and angle of the drone, and the data set is visualized to unify the properties showing the same pattern into one and preprocess the properties showing the outliers. It is expected that the proposed big data-based profile can be used to predict and analyze the optimal movement path of an unmanned aerial vehicle.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.2-4
• /
• 2021

In a distributed machine learning system, sharing gradients was considered safe because it did not share original training data. However, recent studies found that malicious attacker could completely restore the original training data from shared gradients. Gradient Leakage Attack is a technique that restoring original training data by exploiting theses vulnerability. In this study, we present the image transformation method based on Discrete Cosine Transform to defend against the Gradient Leakage Attack on the federated learning setting, which training in local devices and sharing gradients to the server. Experiment shows that our image transformation method cannot be completely restored the original data from Gradient Leakage Attack.

• The Journal of Korea Robotics Society
• /
• v.15 no.2
• /
• pp.184-189
• /
• 2020

This paper is a study on data augmentation for small dataset by using deep learning. In case of training a deep learning model for recognition and classification of non-mainstream objects, there is a limit to obtaining a large amount of training data. Therefore, this paper proposes a data augmentation method using perspective transform and image synthesis. In addition, it is necessary to save the object area for all training data to detect the object area. Thus, we devised a way to augment the data and save object regions at the same time. To verify the performance of the augmented data using the proposed method, an experiment was conducted to compare classification accuracy with the augmented data by the traditional method, and transfer learning was used in model learning. As experimental results, the model trained using the proposed method showed higher accuracy than the model trained using the traditional method.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.4
• /
• pp.270-278
• /
• 2023

We propose a container orchestration system for process workloads that combines the potential of big data and machine learning technologies to integrate enterprise process-centric workloads. This proposed system analyzes big data generated from industrial automation to identify hidden patterns and build a machine learning prediction model. For each machine learning case, training data is loaded into a data store and preprocessed for model training. In the next step, you can use the training data to select and apply an appropriate model. Then evaluate the model using the following test data: This step is called model construction and can be performed in a deployment framework. Additionally, a visual hierarchy is constructed to display prediction results and facilitate big data analysis. In order to implement parallel computing of PCA in the proposed system, several virtual systems were implemented to build the cluster required for the big data cluster. The implementation for evaluation and analysis built the necessary clusters by creating multiple virtual machines in a big data cluster to implement parallel computation of PCA. The proposed system is modeled as layers of individual components that can be connected together. The advantage of a system is that components can be added, replaced, or reused without affecting the rest of the system.

• Journal of International Academy of Physical Therapy Research
• /
• v.6 no.2
• /
• pp.902-907
• /
• 2015

The purpose of this study was to parallel circuit training and circuit training with sonic systemic mechanism was to compare the differences in pulmonary function and chest expansion in adult men. This study was performed on 20 subjects. 20 subjects were divided into two groups; Circuit training group(n=10), Circuit training with sonic systemic mechanism(n=10). Both of the group performed the exercise 3 times a week for 5 weeks. The data was analyzed by the Repeated t-test for comparing before, during and after changes of factors in each group and the Independent t-test for comparing the between groups. The result are as follows. Circuit training group was statistically significant difference FVC, FEV1/FVC(p<.05), Circuit training with sonic systemic mechanism group was statistically significant difference PEF, VC in pulmonary function(p<.05). Circuit training group was statistically significant difference FEV1/FVC of between the two group in pulmonary function(p<.05). Circuit training group and circuit training with sonic systemic mechanism group was statistically significant difference in chest expansion(p<0.05) and there was no statistically significant difference of between the two group in chest expansion(p>.05).

• Journal of Practical Engineering Education
• /
• v.15 no.2
• /
• pp.273-283
• /
• 2023

Education has recently witnessed a rapid increase in the use of edutech worldwide. This study focuses on Korean workers and explores an edutech-based learning model for vocational education and training. Based on analyses of edutech cases and interviews with edutech experts, a draft edutech model was designed and the validity was evaluated based on two Delphi surveys with a panel of experts in the field. The study's findings suggest that edutech-based employee education and training should prioritize LXP orientation (last CVR=1, last Mean=4.70) , implement adaptive learning through learning analytics (last CVR=1, last Mean=4.90), enhance the human touch effect using edutech (last CVR=1, last Mean=4.90), and emphasize the importance of designing curricula that apply edutech in a step-by-step learning process while incorporating suitable instructional design for the key technologies involved in vocational training programs. In addition, it was revealed that there is a strong need to implement a method that makes each stage of the learning process more effective (before, during, and after). Edutech-based vocational training program should consider the interests of all stakeholders, including learners, instructors, vocational training institutions, and government agencies. Given the promotion of government-sponsored vocational training projects in Korea, the findings of this research are likely to have significant implications for the future of Korea's education and training policies.

• Journal of Korean society of Dental Hygiene
• /
• v.24 no.1
• /
• pp.69-77
• /
• 2024

Objectives: The purpose of this study was to explore the general public's perception and status of oral muscle strength training, to develop age-appropriate educational media and training methods, and to promote the need for oral muscle strength training. Methods: Data were collected from 15 individuals across different age groups (young, middle-aged, and elderly) from December 2022 to February 2023 through focus group interviews, and they were conducted twice for each group in a face-to-face manner. Results: Four key categories were identified: lack of information, effectiveness of training, need for promotion, and factors necessary for implementation. The following themes emerged: lack of information, need for training, age-specific characteristics, need for repetition, age at which training is needed, lack of promotion, need for promotion, number of practitioners, willingness to practice, and appropriate media for training. Conclusions: Awareness of oral muscle strength training was found to be very low, and it is necessary to improve awareness through continuous information and appropriate education on its need among the public. Additionally, quality content or media that can be easily applied for effective training should be developed, and personnel who can perform training efficiently should be trained.

• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.95-101
• /
• 1999

In this paper, a comparison between different methods to combine predictions from neural networks will be given. These methods are bagging, bumping, and balancing. Those are based on the analysis of the ensemble generalization error into an ambiguity term and a term incorporating generalization performances of individual networks. Neural Networks and AI machine learning models are prone to overfitting. A strategy to prevent a neural network from overfitting, is to stop training in early stage of the learning process. The complete data set is spilt up into a training set and a validation set. Training is stopped when the error on the validation set starts increasing. The stability of the networks is highly dependent on the division in training and validation set, and also on the random initial weights and the chosen minimization procedure. This causes early stopped networks to be rather unstable: a small change in the data or different initial conditions can produce large changes in the prediction. Therefore, it is advisable to apply the same procedure several times starting from different initial weights. This technique is often referred to as training ensembles of neural networks. In this paper, we presented a comparison of three statistical methods to prevent overfitting of neural network.

• Korean Journal of Adult Nursing
• /
• v.24 no.6
• /
• pp.627-634
• /
• 2012

Purpose: The purpose of this study was to identify registered nurses learning needs about physical assessment. Specifically, what are the perceived competency, frequency of skill use and the unmet training needs. Methods: The study was an exploratory survey study. The sample was 104 registered nurses. Data were collected through three instruments: The Perceived Competency in Physical Assessment Scale, the Frequency of Physical Assessment Scale, and the Training Needs of Physical Assessment Scale which incorporated 30 core Physical Assessment skills. Descriptive statistics, t-test, and Pearson's correlation coefficient were used to analyze the data. Results: Auscultation of heart and lung sounds and inspection of the spine were rated by the subjects as physical assessment skills they feel least competent and also were less frequently performed. The most competent area for physical assessment was neurological system. The respiratory and abdominal system was identified as two systems that more education would be needed. Nurses with less than one year of working experience reported needing more training. Nurses with more than five years of clinical work experience performed physical assessment more frequently than nurses with less than five year of work experience. The perceived competency was positively related to the frequency of physical assessment. Conclusion: Continuing education is necessary to further train registered nurses regarding physical assessment skills and the program needs to be focused on the area which nurses are less competent for and have high training need.