• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.6
• /
• pp.323-331
• /
• 2013

An extreme learning machine (ELM) is an efficient learning algorithm that is based on the generalized single, hidden-layer feed-forward networks (SLFNs), which perform well in classification applications. Many studies have demonstrated its superiority over the existing classical algorithms: support vector machine (SVM) and BP neural network. This paper presents a novel face recognition approach based on a multi-class project extreme learning machine (MPELM) classifier and 2D Gabor transform. First, all face image features were extracted using 2D Gabor filters, and the MPELM classifier was used to determine the final face classification. Two well-known face databases (CMU-PIE and ORL) were used to evaluate the performance. The experimental results showed that the MPELM-based method outperformed the ELM-based method as well as other methods.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.7 no.2
• /
• pp.95-101
• /
• 2012

In order to supply emotion service depending on user's emotional change in a mobile environment, various researches have been carried. This paper discusses a protocol for wireless charging and an embedded platform of the mobile emotional sensing device which supports that. Wireless charging process relieves user's vexatious task to charge the emotional sensing device. To support wireless charging, there are one basestation and several mobile devices. Basestation coordinates and controls the devices over wireless communication, as well as supplies energy. For 1:N communication we defines the network whose superframe is classified into four categories: a network join superframe, a charging request superframe, a charging superframe and an inactive superframe. Physical layer provides how to supply energy to the devices and communicate physically. Mobile device is equipped with energy charged circuits, which correspond with the defined energy supplying method, as well as bidirectional communication circuits. Mobile device monitors and analyzes its own battery status, and is able to send a request packet to basestation. Therefore, it can be charged before its battery is exhausted without user's perception.

• International Journal of Internet, Broadcasting and Communication
• /
• v.12 no.1
• /
• pp.144-150
• /
• 2020

Reinforcement learning is a technology that can present successful and creative solutions in many areas. This reinforcement learning technology was used to deploy containers from cloud servers to fog servers to help them learn the maximization of rewards due to reduced traffic. Leveraging reinforcement learning is aimed at predicting traffic in the network and optimizing traffic-based fog computing network environment for cloud, fog and clients. The reinforcement learning system collects network traffic data from the fog server and IoT. Reinforcement learning neural networks, which use collected traffic data as input values, can consist of Long Short-Term Memory (LSTM) neural networks in network environments that support fog computing, to learn time series data and to predict optimized traffic. Description of the input and output values of the traffic-based reinforcement learning LSTM neural network, the composition of the node, the activation function and error function of the hidden layer, the overfitting method, and the optimization algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.4
• /
• pp.1869-1886
• /
• 2018

This paper proposes a MIDI- and audio-based music genre classification method for Korean traditional music. There are many traditional instruments in Korea, and most of the traditional songs played using the instruments have similar patterns and rhythms. Although music information processing such as music genre classification and audio melody extraction have been studied, most studies have focused on pop, jazz, rock, and other universal genres. There are few studies on Korean traditional music because of the lack of datasets. This paper analyzes raw audio and MIDI phrases in Korean traditional music, performed using Korean traditional musical instruments. The classified samples and MIDI, based on our classification system, will be used to construct a database or to implement our Kontakt-based instrument library. Thus, we can construct a management system for a Korean traditional music library using this classification system. Appropriate feature sets for raw audio and MIDI phrases are proposed and the classification results-based on machine learning algorithms such as support vector machine, multi-layer perception, decision tree, and random forest-are outlined in this paper.

• Journal of the Korea Society of Computer and Information
• /
• v.4 no.1
• /
• pp.47-53
• /
• 1999

In medical field, the computer has been used in the automatic processing of data derived in hospital. the automation of diagonal devices, and processing of medical digital images. In this paper, we classify red blood cell into 16 class including normal cell to the automation of blood analysis to diagnose disease. First, using UNL Fourier and invariant moment algorithm, we extract features of red blood cell from blood cell image and then construct multi-layer backpropagation neural network to recognize. We proof that the system can give support to blood analyzer through blood sample analysis of 10 patients.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.3 no.4
• /
• pp.399-422
• /
• 2009

Although current wireless mesh network (WMN) applications experience a perceptually uninterrupted hand-off, their throughput after the hand-off event may be significantly degraded due to the low available bandwidth of the mobile client's new master. In this paper, we propose a novel mobility management scheme for 802.11-based WMNs that enables both seamless hand-off for transparent communications, and bandwidth awareness for stable application performance after the hand-off process. To facilitate this, we (i) present a new buffer moment in support of the fast Layer-2 hand-off mechanism to cut the packet loss incurred in the hand-off process to zero and (ii) design a dynamic admission control to grant joining accepts to mesh clients. We evaluate the benefits and drawbacks of the proposal for both UDP and TCP traffic, as well as the fairness of the proposal. Our results show that the new scheme can not only minimize hand-off latency, but also maintain the current application rates of roaming users by choosing an appropriate new master for joining.

• Journal of the Korean earth science society
• /
• v.29 no.5
• /
• pp.428-436
• /
• 2008

This study observed the upper tropospheric ozone enhancement in the northern Atlantic for the Aerosols99 campaign in January-February 1999. To find the origin of this air, we have analyzed the horizontal and vertical fields of Isentropic Potential Vorticity (IPV) and Relative Humidity (RH). The arch-shaped IPV is greater than 1.5 pvus indicating stratospheric air stretches equatorward. These arch-shaped regions are connected with regions of RH less than 20%. The vertical fields of IPV and RH show the folding layer penetrating into the upper troposphere. These features support the idea that the upper tropospheric ozone enhancement originated from the stratosphere. Additionally, we have investigated the climatological frequency of stratospheric intrusion over the tropical north Atlantic using IPV and RH. The total frequency between the equator and $30^{\circ}N$ over the tropical north Atlantic exhibits a maximum in northern winter. It suggests that the stratospheric intrusion plays an important role in enhancing ozone in the upper troposphere over the tropical north Atlantic in winter and early spring. Although the tropospheric ozone residual method assumed zonally invariant stratospheric ozone, stratospheric zonal ozone variance could be caused by stratospheric intrusions. This implies that stratospheric intrusion influences ozone variance over the Atlantic in boreal winter and spring, and the intrusion is a possible source for the tropical north Atlantic paradox.

• Transactions of the Korean hydrogen and new energy society
• /
• v.24 no.3
• /
• pp.200-205
• /
• 2013

The separation of hydrogen depends on porosity, diffusivity and solubility in permeation membrane. Dense membrane is always showing a solution diffusion mechanism but porous membrane is not showing. Therefore, porous membrane has a good hydrogen flux due to pore is carried out transferred media. This mechanism is named as the Knudsen diffusion. Hydrogen molecules or hydrogen atoms are diffused along pore that is a mean free path. In this study, complex layer hydrogen permeation membrane was fabricated by hot press process. And then, it was evaluated and calculated to relationship between hydrogen permeability and membrane porosity.

• Journal of Broadcast Engineering
• /
• v.16 no.6
• /
• pp.1073-1076
• /
• 2011

In this paper, we propose a scalable random access scheme in SVC based DASH service that enables random access support not only for base layer of SVC but also for enhancement layers. The proposed method includes extension of segment index box ('sidx') from DASH standard, as well as new RAP Synchronization Box ('raps'). Since the proposed scheme provides random access service for movie fragments with SVC encoded video layers, adaptive scalable random access service is possible.

• Tunnel and Underground Space
• /
• v.3 no.1
• /
• pp.63-79
• /
• 1993

The design of tunnels in rock masses often demands more informations on geologic features and rock mass properties than acquired by usual field survey and laboratory testings. In practice, the situation that a perfect set of geological and mechanical input data is given to geomechanics design engineer is rare, while the engineers are asked to achieve a high level of reliability in their design products. This study presents an artificial neural network which is developed to resolve the difficulties encountered in conventional design techniques, particulary the problem of deteriorating the confidence of existing numerical techniques such as the finite element, boundary element and distinct element methods due to the incomplete adn vague input data. The neural network has inferring capabilities to identify the possible failure modes, support requirements and its timing for underground openings, from previous case histories. Use of the neural network has resulted in a better estimate of the correlation between systems of rock mass classifications such as the RMR and Q systems. A back propagation learning algorithm together with a multi-layer network structure is adopted to enhance the inferential accuracy and efficiency of the neural network. A series of experiments comparing the results of the neural network with the actual field observations are performed to demonstrate the abilities of the artificial neural network as a new tunnel design assistance system.