• Korean Journal of Remote Sensing
• /
• v.38 no.6_4
• /
• pp.1925-1934
• /
• 2022

Monitoring nuclear activity for inaccessible areas using remote sensing technology is essential for nuclear non-proliferation. In recent years, deep learning has been actively used to detect nuclear-activity-related small objects. However, high-resolution satellite imagery containing small objects can result in class imbalance. As a result, there is a performance degradation problem in detecting small objects. Therefore, this study aims to improve detection accuracy by analyzing the effect of the ratio of small objects related to nuclear activity in the input data for the performance of the deep learning model. To this end, six case datasets with different ratios of small object pixels were generated and a U-Net model was trained for each case. Following that, each trained model was evaluated quantitatively and qualitatively using a test dataset containing various types of small object classes. The results of this study confirm that when the ratio of object pixels in the input image is adjusted, small objects related to nuclear activity can be detected efficiently. This study suggests that the performance of deep learning can be improved by adjusting the object pixel ratio of input data in the training dataset.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.323-332
• /
• 2022

As various services using AI technology are being developed, much attention is being paid to AI service production. Recently, AI technology is acknowledged as one of ICT services, a lot of research is being conducted for general-purpose AI service production. In this paper, I describe the research results in terms of systems for AI service production, focusing on the distribution and production of machine learning models, which are the final steps of general machine learning development procedures. Three different Ubuntu systems were built, and experiments were conducted on the system, using data from 2017 validation COCO dataset in combination of different AI models (RFCN, SSD-Mobilenet) and different communication methods (gRPC, REST) to request and perform AI services through Tensorflow serving. Through various experiments, it was found that the type of AI model has a greater influence on AI service inference time than AI machine communication method, and in the case of object detection AI service, the number and complexity of objects in the image are more affected than the file size of the image to be detected. In addition, it was confirmed that if the AI service is performed remotely rather than locally, even if it is a machine with good performance, it takes more time to infer the AI service than if it is performed locally. Through the results of this study, it is expected that system design suitable for service goals, AI model development, and efficient AI service production will be possible.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.5
• /
• pp.517-529
• /
• 2021

This study analyzes changes in the efficiency of the Department of Early Childhood Education in Korea to examine the effectiveness of the National Evaluation for Teacher Education Institution. We provide policy implications from exploring factors influencing efficiency and comparing characteristics of efficient and inefficient departments. With 149 Department of Early Childhood Education in Korea, DEA was conducted to estimate the relative efficiency, and the Tobit model was applied to explore factors affecting efficiency. The results confirm that the Department of Early Childhood in Korea is run efficiently, though there was no change in scale efficiency and relative efficiency after the two phases of the National Evaluation for Teacher Education Institution. The results show the number of books per student was significantly lower despite a significantly higher employment rate. Efficiency of college departments, department greater than 60 (per cohort), and department in metropolitan city (except Seoul area) is confirmed greater. These results provide policy implications on developing evaluation measure and processes to improve education quality and efficiency.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.48 no.2
• /
• pp.129-139
• /
• 2021

The purpose of this study was to analyze the characteristics of pediatric and adolescent patients who received sealant after health insurance coverage based on demographic factors such as gender, age, insurance type, care institution and region. Patient Sample Data from the Health Insurance Review and Assessment Service were used for this study. A total of 8,454,636 patients' data were obtained from 2010 to 2017. Of these, 114,680 patients got sealant treatment. Females were more likely to receive sealant treatment than males. 5 - 9 year age group showed the highest number of patients and proportion of treatment. Patients with health insurance were more likely to receive pit and fissure sealant treatment compared to patients with medical aid program. The number of sealant patients and the proportion of sealant treatment were the highest in dental clinics, followed by dental hospitals and public health centers. The number of sealant patients were the highest in Gyeonggi and proportion of sealant patients were the highest in Jeonbuk.

• Journal of Korea Entertainment Industry Association
• /
• v.14 no.7
• /
• pp.487-500
• /
• 2020

This research analyzed articles about the disability under the 「The Disabled Person Welfare Law」 in a major daily newspaper. A total of 7,684 articles on disability were collected from homepages of the three newspapers , , and . Through network text analysis and content analysis, we considered about "The perspective of Disability" based on "Multiple Disability Model". As a result of this research, when comparing individual models versus social models, individual models have a higher rate 64.31% than social models 35.69%. According to the newspapers, the major perception of Disability is a traditional individual model, which means disability must be solved by individuals. In addition, due to low social and institutional supports, the public's attention and consideration required for the disabled, socially weak people. This research implied that despite the changing times of looking at disability, three newspapers are still staying in the traditional paradigm. Therefore, It is required that viewing a disability from the perspective on disabled people, and a mature awareness that recognizes the diversity of individual needs. The significance of this study can be found in the fact that no attempt has been made to treat the disability perspectivec in newspaper articles as quantitative and qualitative data.

• Journal of the Korea Computer Graphics Society
• /
• v.28 no.1
• /
• pp.1-9
• /
• 2022

Non-small cell lung cancer (NSCLC) accounts for a high proportion of 85% among all lung cancer and has a significantly higher mortality rate (22.7%) compared to other cancers. Therefore, it is very important to predict the prognosis after surgery in patients with non-small cell lung cancer. In this study, the types of preoperative chest CT image patches for non-small cell lung cancer patients with tumor as a region of interest are diversified into five types according to tumor-related information, and performance of single classifier model, ensemble classifier model with soft-voting method, and ensemble classifier model using 3 input channels for combination of three different patches using pre-trained ResNet and EfficientNet CNN networks are analyzed through misclassification cases and Grad-CAM visualization. As a result of the experiment, the ResNet152 single model and the EfficientNet-b7 single model trained on the peritumoral patch showed accuracy of 87.93% and 81.03%, respectively. In addition, ResNet152 ensemble model using the image, peritumoral, and shape-focused intratumoral patches which were placed in each input channels showed stable performance with an accuracy of 87.93%. Also, EfficientNet-b7 ensemble classifier model with soft-voting method using the image and peritumoral patches showed accuracy of 84.48%.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.1
• /
• pp.24-32
• /
• 2023

The current Private 5G use procedure goes through the step of application examination, use and usage inspection, and can be divided in to application, examination step as a procedure before frequency allocation, and use, usage inspection step as a procedure after frequency allocation. Various types of documents are required to apply for a Private 5G, and due to the document screening process and radio station inspection for using Private 5G frequencies, the procedure for Private 5G applicants to use Private 5G is complicated and takes a considerable amount of time. In this paper, we proposed Frequency Allocation Process for Private 5G using a blockchain platform, which is fast and simplified than the current procedure. Through the use of a blockchain platform and NFT (Non-Fungible Token), reliability and integrity of the data required in the frequency allocation process were secured, and security of frequency usage information was maintained and a reliable Private 5G frequency allocation process was established. Also by applying the RPA system that minimizes human intervention, fairness was secured in the process of allocating Private 5G. Finally, the frequency allocation process of Private 5G based on the Ethereum blockchain was performed though a simulation.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.1
• /
• pp.41-49
• /
• 2023

A convolutional neural network (CNN) is a representative algorithm for implementing artificial neural networks. CNNs have improved on the issues of rapid increase in calculation amount and low object classification rates, which are associated with a conventional multi-layered fully-connected neural network (FNN). However, because of the rapid development of IT devices, the maximum resolution of images captured by current smartphone and tablet cameras has reached 108 million pixels (MP). Specifically, a traditional CNN algorithm requires a significant cost and time to learn and process simple, high-resolution images. Therefore, this study proposes an improved CNN algorithm for implementing an object classification learning model for simple, high-resolution images. The proposed method alters the adjacency matrix value of the pooling layer's max pooling operation for the CNN algorithm to reduce the high-resolution image learning model's creation time. This study implemented a learning model capable of processing 4, 8, and 12 MP high-resolution images for each altered matrix value. The performance evaluation result showed that the creation time of the learning model implemented with the proposed algorithm decreased by 36.26% for 12 MP images. Compared to the conventional model, the proposed learning model's object recognition accuracy and loss rate were less than 1%, which is within the acceptable error range. Practical verification is necessary through future studies by implementing a learning model with more varied image types and a larger amount of image data than those used in this study.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.6
• /
• pp.87-96
• /
• 2022

The traffic vulnerable, including elderly pedestrians, have a relatively low walking speed and slow cognitive response time due to reduced physical ability. Although a smart crossing system has been developed and operated to improve problem, it is difficult to operate a signal that reflects the appropriate walking speed for each pedestrian. In this study, a neural network model and a multiple regression model-based traversing speed estimation model were developed using image information collected in an area with a high percentage of traffic vulnerability. to support the provision of optimal walking signals according to real-time traffic weakness. actual traffic data collected from the urban traffic network of Paju-si, Gyeonggi-do were used. The performance of the model was evaluated through seven selected indicators, including correlation coefficient and mean absolute error. The multiple linear regression model had a correlation coefficient of 0.652 and 0.182; the neural network model had a correlation coefficient of 0.823 and 0.105. The neural network model showed higher predictive power.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.225-227
• /
• 2022

Now In Seoul, about 75,000 CCTVs are installed in 25 district offices. Each ward office in Seoul has built a control center for CCTV control and is building information such as people, vehicle types, license plate recognition and color classification into big data through 24-hour artificial intelligence intelligent image analysis. Seoul Metropolitan Government has signed MOUs with the Ministry of Land, Infrastructure and Transport, the National Police Agency, the Fire Service, the Ministry of Justice, and the military base to enable rapid response to emergency/emergency situations. In other words, we are building a smart city that is safe and can prevent disasters by providing CCTV images of each ward office. In this paper, the CCTV image is designed to extract the characteristics of the vehicle and personnel when an incident occurs through artificial intelligence, and based on this, predict the escape route and enable continuous tracking. It is designed so that the AI automatically selects and displays the CCTV image of the route. It is designed to expand the smart city integration platform by providing image information and extracted information to the adjacent ward office when the escape route of a person or vehicle related to an incident is expected to an area other than the relevant jurisdiction. This paper will contribute as basic data to the development of smart city integrated platform research.