• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.4
• /
• pp.1989-2011
• /
• 2019

Vehicle detection based on aerial images is an interesting and challenging research topic. Most of the traditional vehicle detection methods are based on the sliding window search algorithm, but these methods are not sufficient for the extraction of object features, and accompanied with heavy computational costs. Recent studies have shown that convolutional neural network algorithm has made a significant progress in computer vision, especially Faster R-CNN. However, this algorithm mainly detects objects in natural scenes, it is not suitable for detecting small object in aerial view. In this paper, an accurate and effective vehicle detection algorithm based on Faster R-CNN is proposed. Our method fuse a hyperactive feature map network with Eltwise model and Concat model, which is more conducive to the extraction of small object features. Moreover, setting suitable anchor boxes based on the size of the object is used in our model, which also effectively improves the performance of the detection. We evaluate the detection performance of our method on the Munich dataset and our collected dataset, with improvements in accuracy and effectivity compared with other methods. Our model achieves 82.2% in recall rate and 90.2% accuracy rate on Munich dataset, which has increased by 2.5 and 1.3 percentage points respectively over the state-of-the-art methods.

• Journal of Digital Convergence
• /
• v.17 no.7
• /
• pp.341-349
• /
• 2019

Today, art education needs education of media image production or media in the extended field. The purpose of this study is to develop the basic visual grammar and communication ability. We developed and applied the teaching plan for the interview video production as the fusion lesson. 3 classes were conducted for 30 middle school students. The results obtained through the analysis of the pre & post questions and student works are as follows. First, the students used the shot size and stable composition through the interview video production. As a result, this class was effective for the acquisition of the basic video grammar. Second, students improved their ability to ask and listen, understand and organize through interview. As a result, students were able to apply the basic video grammar through the interview video production lesson, effectively producing the interview image.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.7
• /
• pp.2304-2320
• /
• 2021

Intelligently detecting anomalies in health sensor data streams (e.g., Electrocardiogram, ECG) can improve the development of E-health industry. The physiological signals of patients are collected through sensors. Timely diagnosis and treatment save medical resources, promote physical health, and reduce complications. However, it is difficult to automatically classify the ECG data, as the features of ECGs are difficult to extract. And the volume of labeled ECG data is limited, which affects the classification performance. In this paper, we propose a Generative Adversarial Network (GAN)-based deep learning framework (called CAB) for heart arrhythmia classification. CAB focuses on improving the detection accuracy based on a small number of labeled samples. It is trained based on the class-imbalance ECG data. Augmenting ECG data by a GAN model eliminates the impact of data scarcity. After data augmentation, CAB classifies the ECG data by using a Bidirectional Long Short Term Memory Recurrent Neural Network (Bi-LSTM). Experiment results show a better performance of CAB compared with state-of-the-art methods. The overall classification accuracy of CAB is 99.71%. The F1-scores of classifying Normal beats (N), Supraventricular ectopic beats (S), Ventricular ectopic beats (V), Fusion beats (F) and Unclassifiable beats (Q) heartbeats are 99.86%, 97.66%, 99.05%, 98.57% and 99.88%, respectively. Unclassifiable beats (Q) heartbeats are 99.86%, 97.66%, 99.05%, 98.57% and 99.88%, respectively.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.9
• /
• pp.99-112
• /
• 2019

Cracks on pavement surfaces are critical signs and symptoms of the degradation of pavement structures. Image-based pavement crack detection is a challenging problem due to the intensity inhomogeneity, topology complexity, low contrast, and noisy texture background. In this paper, we address the problem of pavement crack detection and segmentation at pixel-level based on a Deep Neural Network (DNN) using gray-scale images. We propose a novel DNN architecture which contains a modified U-net network and a high-level features network. An important contribution of this work is the combination of these networks afforded through the fusion layer. To the best of our knowledge, this is the first paper introducing this combination for pavement crack segmentation and detection problem. The system performance of crack detection and segmentation is enhanced dramatically by using our novel architecture. We thoroughly implement and evaluate our proposed system on two open data sets: the Crack Forest Dataset (CFD) and the AigleRN dataset. Experimental results demonstrate that our system outperforms eight state-of-the-art methods on the same data sets.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3976-3990
• /
• 2022

Currently, cross-attention is widely used in Siamese trackers to replace traditional correlation operations for feature fusion between template and search region. The former can establish a similar relationship between the target and the search region better than the latter for robust visual object tracking. But existing trackers using cross-attention only focus on rich semantic information of high-level features, while ignoring the appearance information contained in low-level features, which makes trackers vulnerable to interference from similar objects. In this paper, we propose a Multi-level Cross-attention Siamese network(MCSiam) to aggregate the semantic information and appearance information at the same time. Specifically, a multi-level cross-attention module is designed to fuse the multi-layer features extracted from the backbone, which integrate different levels of the template and search region features, so that the rich appearance information and semantic information can be used to carry out the tracking task simultaneously. In addition, before cross-attention, a target-aware module is introduced to enhance the target feature and alleviate interference, which makes the multi-level cross-attention module more efficient to fuse the information of the target and the search region. We test the MCSiam on four tracking benchmarks and the result show that the proposed tracker achieves comparable performance to the state-of-the-art trackers.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.8 no.5
• /
• pp.885-892
• /
• 2018

A series of revolutionary industrial changes took place from the 18th century of the First Industrial Revolution. The fourth industrial revolution is a new industrial revolution in which intelligence and information unite. Social, cultural, economic and educational systems are expected to emerge within the category of access and experience. In the course of intelligent mechanization, manpower and machinery need to be commandeered. Tools should be left to the machine and humans should look at essential issues. In the 4th Industrial Revolution, the paradigm of education should fundamentally change. Instead of routine technologies based on memorization, one should learn how to access and utilize. It needs to focus on areas of debate, cooperation, communication, sensibility, and artistry that robots and artificial intelligence can not afford. The fourth industrial revolution is the fusion of human beings and technology, the humanities and the technology.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.6 no.5
• /
• pp.173-185
• /
• 2016

As a dual-degree program/exchange student program becomes widespread, the faculty and students consisting of a course are being globalized. Recently, there has been an online or blended learning-based course, which is evolved from a conventional offline course. However, the course has no consideration of a dual-degree program or exchange student program because it is limited to a domestic university. To create and manage a global course, there have several of difficulties in the aspects of cost or effectiveness. As a solution to tackle them, it is necessary to do research on providing blended global courses with a fusion approach of online and offline. In this paper, we studied and presented a model of blended global courses. To create and maintain overseas scholar-oriented courses, we devised eight types of blended global courses depedning on their opening time and the cooperative relation between online and offline and made their comparative analysis. The blended global course proposed in this paper can be applied to cultivating global human resources in universities.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.344-352
• /
• 2022

Accurate indoor localization of construction workers and mobile assets is essential in safety management. Existing positioning methods based on GPS, wireless, vision, or sensor based RTLS are erroneous or expensive in large-scale indoor environments. Tightly coupled sensor fusion mitigates these limitations. This research paper proposes a state-of-the-art positioning methodology, addressing the existing limitations, by integrating Stereo Visual Inertial Odometry (SVIO) with fiducial landmarks called AprilTags. SVIO determines the relative position of the moving assets or workers from the initial starting point. This relative position is transformed to an absolute position when AprilTag placed at various entry points is decoded. The proposed solution is tested on the NVIDIA ISAAC SIM virtual environment, where the trajectory of the indoor moving forklift is estimated. The results show accurate localization of the moving asset within any indoor or underground environment. The system can be utilized in various use cases to increase productivity and improve safety at construction sites, contributing towards 1) indoor monitoring of man machinery coactivity for collision avoidance and 2) precise real-time knowledge of who is doing what and where.

• Korea Science and Art Forum
• /
• v.20
• /
• pp.147-158
• /
• 2015

This study made an approach to the industrial exhibition space, which is a medium of marketing communication, from the position of an enterprise and consumers through the output of Space Design, and conducted it with focus on B2B transactions among specialized exhibitions. In addition, this study inquired into what factors should be considered along with space design by interpreting the purpose of participating in the exhibition and space design of the enterprise which supply capital goods, elements, related technologies and materials, etc. This study aimed at drawing the direct/indirect effect, produced by space design, on the marketing by analyzing correlation between space design and participating enterprises' marketing. Despite the marketing effect of the exhibition, which was proved by preceding research results, the reality is that exhibition-participating expenses work as considerable burden on enterprises. Particularly, booth design, which is forming the most proportion among the participating expenses, was found to have insufficient influence on visitors due to the decline in its importance among diverse factors influencing visitor's decision to visit a booth. Regardless of the business category of participating enterprises in the exhibition, the standard of exhibits was ranked as the most important consideration factor in visiting a booth. Even by business category, the standard of booth design rarely had an influence on booth visit. Booth design had an affirmative influence on participating enterprise's preference, but its influence on product purchase or business talk & contact with a participating enterprise or price was found to be extremely low. It's difficult to judge marketing success or failure of an exhibition by the form and standard of booth design. Preferably, this study infers that it's necessary to put much weight on qualitative excellence of an exhibition, which consists of participation of an enterprise in possession of excellent technologies, exhibits with higher standards and high-quality visitors with purchasing power. This study suggests that it's more effective to set up the plan for expansion of participation in exhibition by optimally regulating the proportion of space design in participating expense to increase marketing effectiveness of an exhibition. The limitations of this study, analysis of which based on the visitors to an exhibition only, requires supplementation through the follow-up research work on participating enterprises in the exhibition.

• (The) Research of the performance art and culture
• /
• no.38
• /
• pp.249-275
• /
• 2019

Gangneung farm song is important for the ritualistic labortheatricals. Especially, in the theatrical process program, the songsof weeding, such as Haksan Odokkteggi and Ssadae, were alsoincluded to represent the agricultural ritualistic prayer for goodharvest. The use of Gangneung farm song is attractive in termsofthe circumstances. There is no chance of success as a plan simplyfor local events. As a comprehensive art, Gangneung farmsongshould simultaneously revitalize locality, representability andglobality. In this context, three directions are suggested dependingon the development progress. The second direction to the thirddirection are mainly oriented toward government-local authoritycentered support, but, in particular, the third direction should aimtobe an open archive in which the public and the contents prosumercan participate. In the production of the traditional content, OSMUmust be endlessly recreated and spread through 'diversity' and'interrelationships'. The contents utilization was discussed based on the transmissiondetails and analysis evidence of Gangneung Haksan Odokttegi. Thediversification of the way of performing farm song and the folk songs that can induce interest were proposed. Furthermore, basedon the fact that 12 songs classified into 8 work types, there-creation of the contents that focus on the playful aspect of theperforming arts was also proposed. In addition to suggestingthedesirable discourse of the performance groups, the experienceprogram development, education, public relations andcommunication reinforcement based on the narrating principlesoffarm songs, and the development and execution of farmsongprograms based on the tradition and the event demands wereproposed. While mentioning the necessity of value creation, thevitality of performing arts skills that put emotion, imagination, fun, sympathy, and etc. on the original form and the model of farmsongs should be supplemented to Gangneung Haksan Odokttegi. The independent transmission aspect of Gangneung Arirangisalso positive. There are difficulties of reduction of traditionalitydueto the designation, standardized transmission, lack of developmentof voluntary performances, and demonstration-centeredtransmission. Narrowly, we focused on the education andvitalization directions in terms of the capacity building of theconservation council itself. By embodying the essential value of theagricultural literature that farm songs possess, it can bringimpression of all worlds, all lives, communication and sharing, anddraw inspiration from the humanistic view of the world andemotion. The reconstitution of agricultural heritage, theconstruction of a storytelling linked with the concrete objects suchas Beomil National Preceptor, Gulsansa, FlagpoleSupport(Dangganjiju), Seokcheon, Hakbawi rock, and etc., andtheestablishment of fusion type farm song museum were proposed. In terms of culture industry, we requested active support fromtheinstitutions including local governments and various farmsongstorytelling creation projects.