• Convergence Security Journal
• /
• v.20 no.2
• /
• pp.145-153
• /
• 2020

The aviation safety policy master plan is promoting the development of aviation safety management technology applying the 4th industrial revolution technology with the goal of establishing a flawless aviation safety management system and establishing a future aviation safety infrastructure. The master plan includes the establishment of various aviation safety management systems such as aircraft fault management using AI & Big data and flight training system using VR/AR. Currently, the Air Force is promoting a flight safety management system using new technology under the goal of building smart air force. Therefore, this study intends to apply the 4th Industrial Revolution technology to the aircraft condition check system that finally checks the safety of the aircraft before flight. The Air Force conducts airframe flaw checks and pre-flight aircraft check. In this study, we architect the airframe flaw check system using AI and drones, and the pre-flight aircraft condition check system using the IoT and big data for more precise and detailed check of aircraft condition and flawlessness check.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.5
• /
• pp.99-104
• /
• 2016

The walking exercise can alleviate stress and also it can improve health fortheir lifetime. Recent development in Information and Communication Technologies (ICT) has laid the foundation for Internet of Things (IoT) to become the future technology. IoT has many applications in industry automation, security, smart homes and cities, education, health etc. In personal health-care domain, IoT is mainly used for monitoring fitness condition by observing current activity of individual. In this paper, we have proposed a novel IoT based personal wellness care system. Proposed system not only keep track of current fitness level but also provide future activity prediction based on history data along with standard recommendations. Predicted activity helps in motivating the individual to achieve the desired fitness level. Initially, we consider only walking activity for testing purpose and later, other types of activities/exercise will be captured for improved health care support.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.400-403
• /
• 2018

The VR / AR market has grown significantly due to the development of Virtual Reality and Augmented Reality, the core technologies of the Fourth Industrial Revolution. According to a report released by the Korea Science and Engineering Corporation (KISTEP), the global VR / AR market will grow to $ 105 billion by 2022. An important key to the growth of the VR / AR market is user immersion. VR is dependent on technology of hardware such as display and sensor for biometric signal recognition. In order to improve user's immersion feeling, it is important to transmit sensor data to display device more accurately and quickly. In this paper, we consider various sensor hardware dependencies of VR, and compare various correction methods and filtering methods to lower the Motion to Photon (MTP) time that user movement is fully reflected on the display using sensor devices.

• Journal of the Korea Convergence Society
• /
• v.9 no.11
• /
• pp.23-28
• /
• 2018

Recent approaches to Relation Extraction methods mostly tend to be limited to mention level relation extractions. These types of methods, while featuring high performances, can only extract relations limited to a single sentence or so. The inability to extract these kinds of data is a terrible amount of information loss. To tackle this problem this paper presents an Augmented External Memory Neural Network model to enable Global Relation Extraction. the proposed model's Global relation extraction is done by first gathering and analyzing the mention level relation extraction by the Augmented External Memory. Additionally the proposed model shows high level of performances in korean due to the fact it can take the often omitted subjects and objectives into consideration.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.3
• /
• pp.19-28
• /
• 2021

Esophagogastroduodenoscopy is a method commonly used for early diagnosis of upper gastrointestinal lesions. However, 10-20 percent of the gastric lesions are reported to be missed, due to human error. And countries including the US, the UK, and Japan, the World Endoscopy Organization (WEO) suggested guidelines about essential gastrointestinal parts to take pictures of so that all gastric lesions are observed. In this paper, we propose deep learning techniques for classification of anatomical sites, aiming for the system that informs practitioners whether they successfully did the gastroscopy without blind spots. The proposed model uses pre-processing modules and data augmentation techniques suitable for gastroscopy images. Not only does the experiment result with a maximum F1 score of 99.6%, but it also shows a error rate of less than 4% based on the actual data. Given the performance results, we found the model to be explainable with the potential to be utilized in the clinical area.

• Journal of Internet of Things and Convergence
• /
• v.8 no.2
• /
• pp.49-54
• /
• 2022

Immersive contents is content that provides a realistic experience by maximizing the user's five senses, and includes virtual reality, augmented reality, and mixed reality. In order to provide a sense of reality to users in immersive content, it is necessary to provide realistic visual images, hearing, and touch. However, due to the rapid change in the environment for developing immersive content, experts in training human resources are having difficulties in designing the curriculum. In this study, we propose a series of educational courses that use drones to acquire and process real-world measurement data and apply the derived data to VR, AR, and MR to help experts in training immersive content develop talent. The design of training process composes through demand survey and analysis of companies, students, and local communities. This study can be a useful resource for education experts who want to train immersive contents manpower.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.148-154
• /
• 2021

Text-Independent speaker verification needs to extract text-independent speaker embedding to improve generalization performance. However, deep neural networks that depend on training data have the potential to overfit text information instead of learning the speaker information when repeatedly learning from the identical time series. In this paper, to prevent the overfitting, we propose a segment unit shuffling layer that divides and rearranges the input layer or a hidden layer along the time axis, thus mixes the time series information. Since the segment unit shuffling layer can be applied not only to the input layer but also to the hidden layers, it can be used as generalization technique in the hidden layer, which is known to be effective compared to the generalization technique in the input layer, and can be applied simultaneously with data augmentation. In addition, the degree of distortion can be adjusted by adjusting the unit size of the segment. We observe that the performance of text-independent speaker verification is improved compared to the baseline when the proposed segment unit shuffling layer is applied.

• Journal of IKEEE
• /
• v.26 no.4
• /
• pp.568-576
• /
• 2022

Automatic modulation classification(AMC) is a core technique in Software Defined Radio(SDR) platform that enables smart and flexible spectrum sensing and access in a wide frequency band. In this study, we propose a simple yet accurate deep learning-based method that allows AMC for variable-size radio signals. To this end, we design a classification architecture consisting of two Convolutional Neural Network(CNN)-based models, namely main and small models, which were trained on radio signal datasets with two different signal sizes, respectively. Then, for a received signal input with an arbitrary length, modulation classification is performed by augmenting the input samples using a self-replicating padding technique to fit the input layer size of our model. Experiments using the RadioML 2018.01A dataset demonstrated that the proposed method provides higher accuracy than the existing methods in all signal-to-noise ratio(SNR) domains with less computation overhead.

• Journal of the Korea Institute of Building Construction
• /
• v.23 no.3
• /
• pp.273-284
• /
• 2023

This study is centered on the combined use of LiDAR(Light Detection and Ranging) and AR(Augmented Reality) technologies during vertical and horizontal frame measurements in construction projects. The intention is to enhance the quality control procedure, elevate accuracy, and curtail manual labor along with time expenditure. Present methods for accuracy inspection in frame construction often grapple with reliability concerns due to subjective interpretation and the scope for human error. This research recommends the application of LiDAR and AR technologies to counter these issues and augment the efficiency of the inspection process, along with facilitating the dissemination of results. The suggested technique involves the collection of 3D point cloud data of the frame utilizing LiDAR and leveraging this data for checks on construction accuracy. Furthermore, the inspection outcomes are fed into a BIM (Building Information Modeling) model, and the results are visualized via AR. Upon juxtaposing this methodology with the current approach, it is evident that it offers benefits in terms of objective inspection, speed, precise result sharing, and potential enhancements to the overall quality and productivity of construction projects.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.12
• /
• pp.509-516
• /
• 2022

Accurate pose prediction of objects in 3D space is an important visual recognition technique widely used in many applications such as scene understanding in both indoor and outdoor environments, robotic object manipulation, autonomous driving, and augmented reality. Most previous works for object pose estimation have the limitation that they require an exact 3D CAD model for each object. Unlike such previous works, this paper proposes a novel neural network model that can predict the poses of unknown objects based on only their RGB color images without the corresponding 3D CAD models. The proposed model can obtain depth maps required for unknown object pose prediction by using an adaptive depth estimator, AdaBins,. In this paper, we evaluate the usefulness and the performance of the proposed model through experiments using benchmark datasets.