• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.2
• /
• pp.751-770
• /
• 2019

Action recognition is an essential task in computer vision due to the variety of prospective applications, such as security surveillance, machine learning, and human-computer interaction. The availability of more video data than ever before and the lofty performance of deep convolutional neural networks also make it essential for action recognition in video. Unfortunately, limited crafted video features and the scarcity of benchmark datasets make it challenging to address the multi-person action recognition task in video data. In this work, we propose a deep convolutional neural network-based Effective Hybrid Learning (EHL) framework for two-person interaction classification in video data. Our approach exploits a pre-trained network model (the VGG16 from the University of Oxford Visual Geometry Group) and extends the Faster R-CNN (region-based convolutional neural network a state-of-the-art detector for image classification). We broaden a semi-supervised learning method combined with an active learning method to improve overall performance. Numerous types of two-person interactions exist in the real world, which makes this a challenging task. In our experiment, we consider a limited number of actions, such as hugging, fighting, linking arms, talking, and kidnapping in two environment such simple and complex. We show that our trained model with an active semi-supervised learning architecture gradually improves the performance. In a simple environment using an Intelligent Technology Laboratory (ITLab) dataset from Inha University, performance increased to 95.6% accuracy, and in a complex environment, performance reached 81% accuracy. Our method reduces data-labeling time, compared to supervised learning methods, for the ITLab dataset. We also conduct extensive experiment on Human Action Recognition benchmarks such as UT-Interaction dataset, HMDB51 dataset and obtain better performance than state-of-the-art approaches.

• Journal of Internet Computing and Services
• /
• v.25 no.1
• /
• pp.177-188
• /
• 2024

As technology evolves, Internet usage continues to grow, resulting in a geometric increase in network traffic and communication volumes. The network path selection process, which is one of the core elements of the Internet, is becoming more complex and advanced as a result, and it is important to effectively manage and analyze it, and there is a need for a representation and visualization method that can be intuitively understood. To this end, this study designs a framework that analyzes network data using BGP, a network path selection method, and applies it to the cyber common operating picture for situational awareness. After that, we analyze the visualization elements required to visualize the information and conduct an experiment to implement a simple visualization. Based on the data collected and preprocessed in the experiment, the visualization screens implemented help commanders or security personnel to effectively understand the network situation and take command and control.

• Proceedings of the KIEE Conference
• /
• 2008.04a
• /
• pp.65-66
• /
• 2008

A vector DLL represents signal tracking scheme utilizing navigation results, and it has been known that it has better performance than a conventional scalar DLL. This paper discusses the structure and conceptual benefits of the vector DLL, and describes implementation of the vector DLL in a software GPS receive. Also, the benefits of the vector DLL are confirmed by an experiment. Through the experiment, the code tracking accuracy between the vector DLL and a scalar DLL implementation is compared in static environments, and the navigation accuracy is analyzed using GPS signals received from a commercial GPS simulator.

• International journal of advanced smart convergence
• /
• v.6 no.4
• /
• pp.73-79
• /
• 2017

The Convolutional Neural Network (CNN) has shown an excellent performance in computer vision task. Applications of CNN include image classification, object detection in images, autonomous driving, etc. This paper will evaluate the performance of CNN model with ReLU and SELU as activation function. The evaluation will be performed on four different choices of hyperparameter which are initialization method, network configuration, optimization technique, and regularization. We did experiment on each choice of hyperparameter and show how it influences the network convergence and test accuracy. In this experiment, we also discover performance improvement when using SELU as activation function over ReLU.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.4 no.1
• /
• pp.18-22
• /
• 2004

It is difficult to understand a person's emotion, since it is subjective and vague. Therefore, we are proposing a method by which will effectively classify human emotions into two types (that is, single emotion and composition emotion). To verify validity of te proposed method, we conducted two experiments based on general inference and $\alpha$-cut, and compared the experimental results. In the first experiment emotions were classified according to fuzzy inference. On the other hand in the second experiment emotions were classified according to $\alpha$-cut. Our experimental results showed that the classification of emotion based on a- cut was more definite that that based on fuzzy inference.

• IE interfaces
• /
• v.15 no.1
• /
• pp.55-63
• /
• 2002

This paper utilizes the FIDO methodology (Function, Information, Dynamic, Organization modeling) which is an enterprise modeling tool that can describe inter-organizational interaction (specifically between prime and sub contractors in this experiment). FIDO follows the standard IDEFO rules in order to demonstrate how a cascading information flow can evolve from a customer to a prime to a subcontractor in a concurrent manner, in a supply chain environment. Background on these processes is presented, followed with the newly derived process and methodology. This is presented in a supply chain management context, and results from an actual experiment at Motorola utilizing subcontractors that supply custom machine parts is presented and reviewed.

• Journal of the korean Society of Automotive Engineers
• /
• v.12 no.5
• /
• pp.85-96
• /
• 1990

The object of this study is to develop the computer program to predict the transmission loss of a perforated tube muffler with mean flow, and to investigate the influence of porosity and mean flow on the performance of the muffler. The numerical model is made by dividing the muffler into small segments and estimating the transfer matrices for each segment. The computer program is developed for the calculation of the transmission loss of a through-or cross-flow perforated muffler. The experiment is performed for the measurement of the transmission loss and/or the pressure drop for various porosity and flow velocity. From the comparison between computation and experiment, is known that the numerical model agrees well with the experimental result. The effect of porosity and flow velocity on the acoustic performance and the flow resistance of a muffler is presented.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.1 no.1
• /
• pp.1-5
• /
• 2001

In this paper, we propose a fuzzy traffic controller that is able to cope with traffic congestion appropriately. In order to consider such situation as loss of green time caused by spillback of upper crossroad, it imports a degree of traffic congestion of upper roads which vehicles on a crossroad are to proceed to. We constructed the equal-partitioned fuzzy traffic controller that uses the membership functions of the same size and shape, and modified the size and shape, and modified the size and shape of its membership functions by the membership function modification algorithm. In experiment, we compared and analyzed the fixed signal controller, the fuzzy traffic controller with the membership of the same size and shape, and the modified fuzzy traffic controller by using the delay time, the proportion of entered vehicles to occurred vehicles and the proportion of passed vehicles to entered vehicles. As a result of experiment, the modified fuzzy controller showed more enhanced performance than others.

• Proceedings of the KIEE Conference
• /
• 1989.07a
• /
• pp.125-128
• /
• 1989

This paper presents a DC servo motor controller using self-tuning PID algorithm, which can support Multi-processor for the real time processing. Computer simulation as well as experiment using Multi-processor(8088) are implemented with self-tuning PID algorithm. Presented algorithm is used to compare the performance of the controller with that of the classical PID controller through computer simulation and experiment. The result which use the Self-Tuning algorithm show that motor output follows the reference input trajectory fairly well inspite of load disturbances and parameter variations.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.2
• /
• pp.276-279
• /
• 2015

This paper propose a method to reduce the backscatter field of the curved aluminum plate using the cancellation system. The cancellation circuit is composed of a circulator, a LNA(Low Noise Amplifier), a VGA(Variable Gain Amplifier) and two phase shifters. Prior to experiment, we performed simulations to confirm the possibility using FDTD(Finite Difference Time Domain) simulator. We confirmed that the backscatter field could be reduced by the cancellation circuit when we changed the appropriate gain and phase. Finally, we performed an experiment to verify the performance of the cancellation circuit.