• Journal of Information Processing Systems
• /
• v.13 no.5
• /
• pp.1203-1212
• /
• 2017

Intrusion detection systems (IDSs) are crucial in this overwhelming increase of attacks on the computing infrastructure. It intelligently detects malicious and predicts future attack patterns based on the classification analysis using machine learning and data mining techniques. This paper is devoted to thoroughly evaluate classifier ensembles for IDSs in IEEE 802.11 wireless network. Two ensemble techniques, i.e. voting and stacking are employed to combine the three base classifiers, i.e. decision tree (DT), random forest (RF), and support vector machine (SVM). We use area under ROC curve (AUC) value as a performance metric. Finally, we conduct two statistical significance tests to evaluate the performance differences among classifiers.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.756-766
• /
• 2023

To effectively track the illegal use of digital images and maintain the security of digital image communication on the Internet, this paper proposes a reversible multipurpose image watermarking algorithm based on a deep residual network (ResNet) and perceptual hashing (also called MWR). The algorithm first combines perceptual image hashing to generate a digital fingerprint that depends on the user's identity information and image characteristics. Then it embeds the removable visible watermark and digital fingerprint in two different regions of the orthogonal separation of the image. The embedding strength of the digital fingerprint is computed using ResNet. Because of the embedding of the removable visible watermark, the conflict between the copyright notice and the user's browsing is balanced. Moreover, image authentication and traitor tracking are realized through digital fingerprint insertion. The experiments show that the scheme has good visual transparency and watermark visibility. The use of chaotic mapping in the visible watermark insertion process enhances the security of the multipurpose watermark scheme, and unauthorized users without correct keys cannot effectively remove the visible watermark.

• Proceedings of the KIEE Conference
• /
• summer
• /
• pp.85-87
• /
• 2004

Supervisory control and data acquisition (SCADA) systems are essential parts of power system which employ a wide range of computers and communication technologies. The traditional SCADA system is mainly for information exchange in only one company, and the information is provided only to the operator or administrator. But in the deregulated environment, we need much more information, which can be exchanged among different companies. With the rapid development of internet, we can use it to access information easily. This paper proposes web technologies to be applied in power system in order to display some important information through accessing data from database, and to realize the real time control of the substation. The functions of SCADA system will be implemented by a set of Web-based components. The monitoring and control of standard 154[kV] substation model is already realized in the laboratory test. The Web-based SCADA system is able to provide sufficient information and control for pow or system through an efficient and economical way.

• The Journal of Korean Institute for Practical Engineering Education
• /
• v.1 no.1
• /
• pp.99-105
• /
• 2009

Computer-based learning with the access to World Wide Web has become a fundamental base for adopting beneficial education. It provides significant facilities such as animation and interactive processes that are not possible with textbooks. Web/Internet-enabled applications which is fully controlled and monitored from remote locations are extensively used by a number of Universities, national laboratories and companies for different kinds of applications all over the world. Continuous advances in computers and electronics coupled with drooping prices of hardware have made Web/Internet-based technologies less costly than before, particularly for educational organizations. Thus, it is more affordable to invest in these technologies that are essential for both expanding education over Web and further improving and advancing such technologies the application of remote virtual instruments will be demonstrated in this context along with experiments that can be adopted to be educational experimental lab for Engineering Education students.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.1
• /
• pp.77-92
• /
• 2020

When automatic speech recognition (ASR) is provided as a cloud service, it is easy to collect voice and application domain data from users. Harnessing these data will facilitate the provision of more personalized services. In this paper, we demonstrate our transfer learning-based knowledge service that built with the user-generated data collected through our novel system that deliveries personalized ASR service. First, we discuss the motivation, challenges, and prospects of building up such a knowledge-based service-oriented system. Second, we present a Quadruple Transfer Learning (QTL) method that can learn a classification model from a source domain and transfer it to a target domain. Third, we provide an overview architecture of our novel system that collects voice data from mobile users, labels the data via crowdsourcing, utilises these collected user-generated data to train different machine learning models, and delivers the personalised real-time cloud services. Finally, we use the E-Book data collected from our system to train classification models and apply them in the smart TV domain, and the experimental results show that our QTL method is effective in two classification tasks, which confirms that the knowledge transfer provides a value-added service for the upper-layer mobile applications in different domains.

• Smart Structures and Systems
• /
• v.26 no.5
• /
• pp.559-573
• /
• 2020

One of the powerful data management tools is Building Information Modeling (BIM) which operates through obtaining, recalling, sharing, sorting and sorting data and supplying a digital environment of them. Employing SHM, a BIM in monitoring systems, would be an efficient method to address their data management problems and consequently optimize the economic aspects of buildings. The recording of SHM data is an effective way for engineers, facility managers and owners which make the BIM dynamic through the provision of updated information regarding the occurring state and health of different sections of the building. On the other hand, digital transformation is a continuous challenge in construction. In a cloud-based BIM platform, environmental and localization data are integrated which shape the Internet-of-Things (IoT) method. In order to improve work productivity, living comfort, and entertainment, the IoT has been growingly utilized in several products (such as wearables, smart homes). However, investigations confronting the integration of these two technologies (BIM and IoT) remain inadequate and solely focus upon the automatic transmission of sensor information to BIM models. Therefore, in this composition, the use of BIM based on SHM and IOT is reviewed and the economic application is considered.

• Journal of Internet of Things and Convergence
• /
• v.8 no.2
• /
• pp.1-6
• /
• 2022

This paper presents a regionalized time slotted channel hopping (TSCH) slotframe-based aerial data collection using wake-up radio. The proposed scheme aims to minimize the delay and energy consumption when an unmanned aerial vehicle (UAV) collects data from sensor devices in the large-scale service area. To this end, the proposed scheme divides the service area into multiple regions, and determines the TSCH slotframe length for each region according to the number of cells required by sensor devices in each region. Then, it allocates the cells dedicated for data transmission to the TSCH slotframe using the ID of each sensor device. For energy-efficient data collection, the sensor devices use a wake-up radio. Specifically, the sensor devices use a wake-up radio to activate a network interface only in the cells allocated for beacon reception and data transmission. The simulation results showed that the proposed scheme exhibited better performance in terms of delay and energy consumption compared to the existing scheme.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.12
• /
• pp.3383-3397
• /
• 2023

Scene graphs serve as semantic abstractions of images and play a crucial role in enhancing visual comprehension and reasoning. However, the performance of Scene Graph Generation is often compromised when working with biased data in real-world situations. While many existing systems focus on a single stage of learning for both feature extraction and classification, some employ Class-Balancing strategies, such as Re-weighting, Data Resampling, and Transfer Learning from head to tail. In this paper, we propose a novel approach that decouples the feature extraction and classification phases of the scene graph generation process. For feature extraction, we leverage a transformer-based architecture and design an adaptive calibration function specifically for predicate classification. This function enables us to dynamically adjust the classification scores for each predicate category. Additionally, we introduce a Distribution Alignment technique that effectively balances the class distribution after the feature extraction phase reaches a stable state, thereby facilitating the retraining of the classification head. Importantly, our Distribution Alignment strategy is model-independent and does not require additional supervision, making it applicable to a wide range of SGG models. Using the scene graph diagnostic toolkit on Visual Genome and several popular models, we achieved significant improvements over the previous state-of-the-art methods with our model. Compared to the TDE model, our model improved mR@100 by 70.5% for PredCls, by 84.0% for SGCls, and by 97.6% for SGDet tasks.

• Nuclear Engineering and Technology
• /
• v.52 no.7
• /
• pp.1565-1570
• /
• 2020

Photons with the energy of 60 keV are regularly used for some kinds of bone density examination devices, like the single photon absorptiometry (SPA). This article reports a flexible and lightweight rare earth/polymer composite for enhancing shielding efficiency against photon radiation with the energy of 60 keV. Lead oxide (PbO) and several rare earth element oxides (La₂O₃, Ce₂O₃, Nd₂O₃) were dispersed into natural rubber (NR) and the photon radiation shielding performance of the composites were assessed using monte carlo simulation method. For 60 keV photons, the shielding efficiency of rare earthbased composites were found to be much higher than that of the traditional lead-based composite, which has bad absorbing ability for photons with energies between 40 keV and 88 keV. In comparison with the lead oxide based composite, Nd₂O₃-NR composite with the same protection standard (the lead equivalent is 0.25 mmPb, 0.35 mmPb and 0.5 mmPb, respectively), can reduce the thickness by 35.29%, 37.5% and 38.24%, and reduce the weight by 38.91%, 40.99% and 41.69%, respectively. Thus, a flexible, lightweight and lead-free rare earth/NR composite could be designed, offering efficient photon radiation protection for the users of the single photon absorptiometry (SPA) with certain energy of 60 keV.

• Journal of Broadcast Engineering
• /
• v.27 no.4
• /
• pp.527-537
• /
• 2022

As multimedia contents demand and supply increase, internet traffic around the world increases. Several standardization groups are striving to establish more efficient compression standards to mitigate the problem. In particular, research to introduce deep learning technology into compression standards is actively underway. Despite the fact that deep learning-based technologies show high performance, they suffer from the domain gap problem when test video sequences have different characteristics of training video sequences. To this end, several methods have been made to introduce content-adaptive deep video compression. In this paper, we will look into these methods by three aspects: codec information-aware methods, model selection methods, and information signaling methods.