• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.5
• /
• pp.53-63
• /
• 2019

To provide a hand gesture interface through deep learning in mobile environments, research on the light-weighting of networks is essential for high recognition rates while at the same time preventing degradation of execution speed. This paper proposes a method of real-time recognition of written characters in the air using a finger on mobile devices through the light-weighting of deep-learning model. Based on the SSD (Single Shot Detector), which is an object detection model that utilizes MobileNet as a feature extractor, it detects index finger and generates a result text image by following fingertip path. Then, the image is sent to the server to recognize the characters based on the learned OCR model. To verify our method, 12 users tested 1,000 words using a GALAXY S10+ and recognized their finger with an average accuracy of 88.6%, indicating that recognized text was printed within 124 ms and could be used in real-time. Results of this research can be used to send simple text messages, memos, and air signatures using a finger in mobile environments.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.42 no.1
• /
• pp.115-128
• /
• 2019

Small and Medium-sized Enterprises (SMEs) are being faced with rapid changes in their business environments due to evolution of technologies and innovation in societal eco-systems. Particularly, dynamic interactions between such environments and enterprise activities have become significant, so technology planning, which is a process of identifying appropriate directions regarding product and technology development, has received much attention to cope with such dynamics proactively. However, SMEs typically have limits in performing independent, strategical and systematical technology planning activities due to the lack of human, material and financial resources. This paper proposes the development of a product roadmapping method so that SMEs carry out efficient technology planning activities with interconnections of external business environments. The present work provides product roadmap templates that directly accommodate the influence of business environments on the product's system and its associated super/sub-systems with the use of external environment analysis techniques including TRIZ methodology, PEST and 5Forces analysis. These templates are useful to efficiently forecast the directions of product's development and evolution, which arise from changes in external environments. Consequently, the present work enables SMEs to flexibly cope with the era of the next R&D generation, which pursues value creation through mutual interconnection between business environments and technology development.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.66-74
• /
• 2018

Recently, there has been active studies on hand gesture recognition to increase immersion and provide user-friendly interaction in a virtual reality environment. However, most studies require specialized sensors or equipment, or show low recognition rates. This paper proposes a hand gesture recognition method using Deep Learning technology without separate sensors or equipment other than camera to recognize static and dynamic hand gestures. First, a series of hand gesture input images are converted into high-frequency images, then each of the hand gestures RGB images and their high-frequency images is learned through the DenseNet three-dimensional Convolutional Neural Network. Experimental results on 6 static hand gestures and 9 dynamic hand gestures showed an average of 92.6% recognition rate and increased 4.6% compared to previous DenseNet. The 3D defense game was implemented to verify the results of our study, and an average speed of 30 ms of gesture recognition was found to be available as a real-time user interface for virtual reality applications.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.75-86
• /
• 2018

The DDS (Data Distribution Service) specification defines a discovery method for finding participants and endpoints in a DDS network. The standard discovery mechanism uses the multicast protocol and finds all the endpoints in the network. Because of using multicasting, discovery may fail in a network with different segments. Other problems include that memory space wastes due to storing information of all the endpoints. The Topic Name Service (TNS) solves these problems by unicasting only the endpoints, which are required for communication. However, an extra delay time is inevitable in components of TNS, i.e, a front-end server, topic name servers, and a terminal server. In this paper, we analyze the performance of TNS. Delay times in the servers of TNS and time required to receive endpoint information are measured. Time to finish discovery and number of receiving endpoints compare with the standard discovery method.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.5
• /
• pp.7-18
• /
• 2018

Recently, serverless computing is spotlighted. Because it supports the development of application based on micro-service. Micro-service means a small-scale service that can operate independently. Applications with micro-service units have the advantage of enabling individual updates, easy and fast deployment. In addition, it has the advantage of supporting various languages and platforms for each service. Therefore many enterprise are trying to change from monolithic architecture to micro-service based architecture. However, there is a lack of research on methods and baseline for micro-service construction. In this paper, a method is proposed to construct the micro-service unit by analyzing UML design in monolithic application. It also shows the proposed approach can reconstruct monolithic application into micro-service based unit by implementing the constructed micro-services in a real serverless platform environment. In addition, the results of the comparative evaluation with the related studies are presented.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.5
• /
• pp.53-59
• /
• 2018

Due to the recent developments in computer vision technology, complex actions can be recognized with reasonable accuracy in smart cities. In contrast, violence recognition such as events related to fight and knife, has gained less attention. The capability of visual surveillance can be used for detecting fight in streets or in prison centers. In this paper, we proposed a deep learning-based violence recognition method for surveillance cameras. A convolutional neural network (CNN) model is trained and fine-tuned on available benchmark datasets of fights and knives for violence recognition. When an abnormal event is detected, an alarm can be sent to the nearest police station to take immediate action. Moreover, when the probabilities of fight and knife classes are predicted very low, this situation is considered as normal situation. The experimental results of the proposed method outperformed other state-of-the-art CNN models with high margin by achieving maximum 99.21% accuracy.

• Journal of Advanced Navigation Technology
• /
• v.26 no.3
• /
• pp.144-160
• /
• 2022

Research on urban air mobility (UAM) is being actively conducted as a method of next-generation transportation. eVTOL, an airplane to be used for urban air mobility, is classified into a complex type, a tilt rotor type, a tilt wing type, a tilt duct fan type, and a multicopter type according to the propulsion method. In this study, conceptual design was performed for the next generation eVTOL of the new tilt rotor type in accordance with the existing design requirements. The aerodynamic analysis programs of OpenVSP and XFLR5 were used to perform aerodynamic analysis. The power required for each flight mission stage was calculated, the battery and motor were selected accordingly, and MTOW (Maximum Take-Off Weight) was predicted by estimating the weight of each component.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.13 no.5
• /
• pp.39-50
• /
• 2017

Moving object tracking techniques can be considered as one of the most essential technique in the video understanding of which the importance is much more emphasized recently. However, irregularity of light condition in the video, variations in shape and size of object, camera motion, and occlusion make it difficult to tracking moving object in the video. Swarm based methods are developed to improve the performance of Kalman filter and particle filter which are known as the most representative conventional methods, but these methods also need to consider dynamic property of moving object. This paper proposes adaptive parameter control method which can dynamically change weight value among parameters in particle swarm optimization. The proposed method classifies each particle to 3 groups, and assigns different weight values to improve object tracking performance. Experimental results show that our scheme shows considerable improvement of performance in tracking objects which have nonlinear movements such as occlusion or unexpected movement.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.13 no.4
• /
• pp.7-18
• /
• 2017

Electrical energy is one of the most important energy sources in modern society. Therefore, it is very important to control the supply and demand of electric power. However, the power consumption data needed to predict power demand may include the information about the private behavior of an individual, the analysis of which may raise privacy issues. In this paper, we propose a secure power demand forecasting method where regression analyses on power consumption data are conducted in a trusted execution environment provided by Intel SGX, keeping the power usage pattern of users private. We performed experiments using various regression equations and selected an equation which has the least error rate. We show that the average error rate of the proposed method is lower than those of the previous forecasting methods with privacy protection functionality.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.13 no.5
• /
• pp.51-70
• /
• 2017

The reliability of broadcasting over vehicular ad hoc networks (VANETs) is one of the most critical factors for driving safety applications. There exists limitations to improve the reliability of broadcast transmissions in saturated VANETs where previous proposals in literature tackle the problem by heuristically adapting the size of the contention window (CW). This paper considers improving the reliability by proposing a new probability model based on the counting methods of permutations and combinations, which counts all the possible cases of broadcast failures in a single-hop broadcast transmission for a given CW. From the model, we calculate the best CW size given the number of contention nodes, which significantly improves the reliability and satisfying the timely dissemination of emergency broadcasting messages. Through extensive VANET simulations with varying densities, we show that our model maintains near 100 percent success rate for single-hop broadcast as well as multi-hop broadcast (e.g. 40 hops) and achieves minimal broadcast delay.