• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2013.05a
• /
• pp.10-11
• /
• 2013

Green Frame is a column-beam system formed by composite precast concrete column and beam connected with steel buried in both members. During the installation of Green columns, the columns of Green Frame, covering 3 floors per each piece and beams, the eccentricity can be observed due to the construction error and the weight of beam itself. Such eccentricity may have a little influence on a single frame, yet, it can develop critical issues to the installation of subsequent beams or beams on the upper floors in the context of a building as a whole that has multiple frames. These issues lead to delay in frame installation, decrease of productivity and increase of cost, etc. Therefore, this study presents a steel-joint connection method in order to solve the issues. The steel-joint connection method exists on slope plane and reinforcing plate in steel frame buried in composite PC members. Through this method, the issues can be resolved without requiring additional equipment or manpower.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38C no.9
• /
• pp.781-786
• /
• 2013

Multiplexing information over parallel data channels based on RF MIMO concept is possible to achieve considerable data rates over large transmission ranges with just a single transmitting element. Visual multiplexing MIMO techniques will send independent streams of bits using the multiple elements of the light transmitter array and recording over a group of camera pixels can further enhance the data rates. The proposed system is a combination of the reliance on computer vision algorithms for tracking and OOK cell frame modulation. LED array are controlled to transmit message in the form of digital information using ON-OFF signaling with ON-OFF pulses (ON = bit 1, OFF = bit 0). A camera captures image frames of the array which are then individually processed and sequentially decoded to retrieve data. To demodulated data transmission, a motion tracking algorithm is implemented in OpenCV (Open source Computer Vision library) to classify the transmission pattern. One of the most advantages of proposed architecture is Computer Vision (CV) based image analysis techniques which can be used to spatially separate signals and remove interferences from ambient light. It will be the future challenges and opportunities for mobile communication networking research.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.12A
• /
• pp.945-950
• /
• 2003

The link aggregation between 10GbE switches requires an advanced frame distribution method to be properly and efficiently applied. The fixed or dynamic frame distribution methods, formerly proposed, cannot fully utilize the aggregated links, where the receiving terminal only attaches to a pre-specified link among multiple physical links. A frame distribution method using tagging is proposed for the link aggregation between 10GbE switches to solve this problem. We compared the performance of the proposed method with those of the fixed and dynamic frame distribution methods. As a result, the proposed method shows a better performance when the applied load is below 0.7 and the average length of the frames is longer than 954 bytes.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.7B
• /
• pp.827-833
• /
• 2011

Directional MAC protocols have drawn great attentions recently in super high speed wireless local/personal area networks due to their higher antenna gain, better spatial reuse, longer transmission range, and lower interference. Despite of these merits, directional MAC protocols suffer from deafness problem. The deafness problem occurs if a node does not answer an RTS frame addressed to it. To overcome this problem, directional MAC protocols have aimed at avoiding the deafness problem using multiple control frames or advance notice techniques or distinguishing deafness from collision. In this paper, we analyze the performance of these schemes in deafness environments with some scenarios. Through performance analysis, we compare the performance of these schemes through evaluating actual network throughput.

• Journal of Information Processing Systems
• /
• v.10 no.2
• /
• pp.176-192
• /
• 2014

The realization of Wireless Multimedia Sensor Networks (WMSNs) has been fostered by the availability of low cost and low power CMOS devices. However, the transmission of bulk video data requires adequate bandwidth, which cannot be promised by single path communication on an intrinsically low resourced sensor network. Moreover, the distortion or artifacts in the video data and the adherence to delay threshold adds to the challenge. In this paper, we propose a two stage Quality of Service (QoS) guaranteeing scheme called Prioritized Multipath WMSN (PMW) for transmitting H.264 encoded video. Multipath selection based on QoS metrics is done in the first stage, while the second stage further prioritizes the paths for sending H.264 encoded video frames on the best available path. PMW uses two composite metrics that are comprised of hop-count, path energy, BER, and end-to-end delay. A color-coded assisted network maintenance and failure recovery scheme has also been proposed using (a) smart greedy mode, (b) walking back mode, and (c) path switchover. Moreover, feedback controlled adaptive video encoding can smartly tune the encoding parameters based on the perceived video quality. Computer simulation using OPNET validates that the proposed scheme significantly outperforms the conventional approaches on human eye perception and delay.

• Structural Engineering and Mechanics
• /
• v.76 no.3
• /
• pp.337-354
• /
• 2020

Reinforced concrete beam-column (RCBC) joints of laterally loaded unbraced frames are sometimes controlled by their shear behavior. This behavior relies on multiple and interdependent complex mechanisms. There are already several studies on the influence of some parameters on the shear strength of reinforced concrete joints. However, there are no studies methodically tackling all the most relevant parameters and quantifying their influence on the overall joint behavior, not just on its shear strength. Hence, considering the prohibitive cost of a comprehensive parametric experimental investigation, a nonlinear finite element analysis (NLFEA) was undertaken to identify the key factors affecting the shear behavior of such joints and quantify their influence. The paper presents and discusses the models employed in this NLFEA and the procedure used to deduce the joint behavior from the NLFEA results. Three alternative, or complementary, quantities related to shear are considered when comparing results, namely, the maximum shear stress supported by the joint, the secant shear stiffness at maximum shear stress and the secant shear stiffness in service conditions. Depending on which of these is considered, the lower or higher the relevance of each of the six parameters investigated: transverse reinforcement in the joint, intermediate longitudinal bars and diagonal bars in the column, concrete strength, column axial load and confining elements in transverse direction.

• Journal of Aerospace System Engineering
• /
• v.13 no.5
• /
• pp.39-48
• /
• 2019

The turbulent flow in the rotor-stator mixer is based on shear characteristics generated by the interaction of the stator with the rotor rotating at high speed. In this study, the flow characteristics analysis of the unsteady state generated by the interaction of the rotor and the stator in the prototype model of the emulsion-fuel related mixer development was performed with the MRF and SMM by applying the ANSYS FLUENT $k-{\varepsilon}$ (RKE) turbulence model. The behavior and shear characteristics of the flow particles generated at the interface between the designed rotor and stator, and trends such as velocity distribution and turbulence eddy dissipation, were predicted and verified using the CFD analysis.

• Journal of Korea Multimedia Society
• /
• v.24 no.10
• /
• pp.1380-1390
• /
• 2021

The video face recognition (FR) is one of the most popular researches in the field of computer vision due to a variety of applications. In particular, research using the attention mechanism is being actively conducted. In video face recognition, attention represents where to focus on by using the input value of the whole or a specific region, or which frame to focus on when there are many frames. In this paper, we propose a novel attention based deep learning method. Main novelties of our method are (1) the use of combining two loss functions, namely weighted Softmax loss function and a Triplet loss function and (2) the feasibility of end-to-end learning which includes the feature embedding network and attention weight computation. The feature embedding network has a positive effect on the attention weight computation by using combined loss function and end-to-end learning. To demonstrate the effectiveness of our proposed method, extensive and comparative experiments have been carried out to evaluate our method on IJB-A dataset with their standard evaluation protocols. Our proposed method represented better or comparable recognition rate compared to other state-of-the-art video FR methods.

• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.337-351
• /
• 2021

Automatically recognizing facial expressions in video sequences is a challenging task because there is little direct correlation between facial features and subjective emotions in video. To overcome the problem, a video facial expression recognition method using spatiotemporal recurrent neural network and feature fusion is proposed. Firstly, the video is preprocessed. Then, the double-layer cascade structure is used to detect a face in a video image. In addition, two deep convolutional neural networks are used to extract the time-domain and airspace facial features in the video. The spatial convolutional neural network is used to extract the spatial information features from each frame of the static expression images in the video. The temporal convolutional neural network is used to extract the dynamic information features from the optical flow information from multiple frames of expression images in the video. A multiplication fusion is performed with the spatiotemporal features learned by the two deep convolutional neural networks. Finally, the fused features are input to the support vector machine to realize the facial expression classification task. The experimental results on cNTERFACE, RML, and AFEW6.0 datasets show that the recognition rates obtained by the proposed method are as high as 88.67%, 70.32%, and 63.84%, respectively. Comparative experiments show that the proposed method obtains higher recognition accuracy than other recently reported methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.8
• /
• pp.4026-4040
• /
• 2019

In this paper, we present an approach for detecting slide transitions in lecture videos by introducing the spatio-temporal residual networks. Given a lecture video which records the digital slides, the speaker, and the audience by multiple cameras, our goal is to find keyframes where slide content changes. Since temporal dependency among video frames is important for detecting slide changes, 3D Convolutional Networks has been regarded as an efficient approach to learn the spatio-temporal features in videos. However, 3D ConvNet will cost much training time and need lots of memory. Hence, we utilize ResNet to ease the training of network, which is easy to optimize. Consequently, we present a novel ConvNet architecture based on 3D ConvNet and ResNet for slide transition detection in lecture videos. Experimental results show that the proposed novel ConvNet architecture achieves the better accuracy than other slide progression detection approaches.