• Proceedings of the IEEK Conference
• /
• 2002.07b
• /
• pp.802-805
• /
• 2002

In the three-dimensional domain image expressed with two-dimensional slice images, such as fMRI images and multi-slice CT images, we propose the three-dimensional domain automatic segmentation for the purpose of extracting region. In this paper, we segmented each domain from the fMRI images of the head of people and monkey. We used the neural network "Pulse-Coupled Neural Network" which is one of the models of visual cortex of the brain based on the knowledge from neurophysiology as the technique. By using this technique, we can segment the region without any learning. Then, we reported the result of division of each domain and extraction to the fMRI slice images of human's head using "three-dimensional Pulse-Coupled Neural Network" which is arranged and created the neuron in the shape of a three-dimensional lattice.

• Journal of Korea Multimedia Society
• /
• v.23 no.12
• /
• pp.1540-1551
• /
• 2020

Aiming at the problem that the existing human behavior recognition algorithm cannot fully utilize the multi-level spatio-temporal information of the network, a human behavior recognition algorithm based on a dense three-dimensional residual network is proposed. First, the proposed algorithm uses a dense block of three-dimensional residuals as the basic module of the network. The module extracts the hierarchical features of human behavior through densely connected convolutional layers; Secondly, the local feature aggregation adaptive method is used to learn the local dense features of human behavior; Then, the residual connection module is applied to promote the flow of feature information and reduced the difficulty of training; Finally, the multi-layer local feature extraction of the network is realized by cascading multiple three-dimensional residual dense blocks, and use the global feature aggregation adaptive method to learn the features of all network layers to realize human behavior recognition. A large number of experimental results on benchmark datasets KTH show that the recognition rate (top-l accuracy) of the proposed algorithm reaches 93.52%. Compared with the three-dimensional convolutional neural network (C3D) algorithm, it has improved by 3.93 percentage points. The proposed algorithm framework has good robustness and transfer learning ability, and can effectively handle a variety of video behavior recognition tasks.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.26 no.10
• /
• pp.1341-1347
• /
• 2002

In order to improve efficiency of a system with three-dimensional flow characteristics, this paper presents a new method that overcomes three-dimensional effects by using two-dimensional CFD and neural network. The method was applied to shape optimization of cut-off in a multi-blade fan/scroll system. As the entrance conditions of two-dimensional CFD, the experimental values at the positions out of the inactive zone were used. The distributions of velocity and pressure obtained by two-dimensional CFD were compared with those of three-dimensional CFD and experimental results. It was found that the distributions of velocity and pressure have qualitative similarity. The results of two-dimensional CFD were used for teaming as target values of neural network. The optimal angle and radius of cut-off were determined as 71$^{\circ}$and 0.092 times the outer diameter of impeller, respectively. It is quantified in the previous report that the optimal angle and radius of cut-off are approximately 72$^{\circ}$and 0.08 times the outer diameter of impeller, respectively.

• Journal of Communications and Networks
• /
• v.11 no.3
• /
• pp.287-296
• /
• 2009

Although asymptotic bounds of wireless network capacity have been heavily pursued, the answers to the following questions are still critical for network planning, protocol and architecture design: Given a three-dimensional (3D) network space with the number of active users randomly located in the space and using the wireless communication technology, what are the expected per-flow throughput, network capacity, and network transport capacity? In addition, how can the protocol parameters be tuned to enhance network performance? In this paper, we focus on the ultra wideband (UWB) based wireless personal area networks (WPANs) and provide answers to these questions, considering the salient features of UWB communications, i.e., low transmission/interference power level, accurate ranging capability, etc. Specifically, we demonstrate how to explore the spatial multiplexing gain of UWB networks by allowing appropriate concurrent transmissions. Given 3D space and the number of active users, we derive the expected number of concurrent transmissions, network capacity and transport capacity of the UWB network. The results reveal the main factors affecting network (transport) capacity, and how to determine the best protocol parameters, e.g., exclusive region size, in order to maximize the capacity. Extensive simulation results are given to validate the analytical results.

• Korea-Australia Rheology Journal
• /
• v.15 no.1
• /
• pp.27-33
• /
• 2003

Associating polymers are hydrophilic long-chain molecules to which a small amount of hydrophobic groups (hydrophobes) is incorporated. In aqueous solution, the association interactions result in the formation of three-dimensional network through flowerlike micelles at high concentrations. In colloidal suspensions, the associating polymers act as flocculated by bridging mechanism. The rheological properties of suspensions flocculated by associating polymers end-capped with hydrophobes are studied in relation to the bridging conformation. At low polymer concentrations, the polymer chains effectively form bridges between particles by multichain association. The suspensions are highly flocculated and show typical viscoelastic responses. When the polymer concentration is increased above the absorbance at saturation, the excess polymer chains remaining in the solution phase build up three-dimensional network by associating interactions. Since the presence of particles does not significantly influence the network structures in the medium, the relative viscosity, which gives a measure of the degree of flocculation is decreased with increasing polymer concentration. The bridging conformation and flocculation level vary strongly depending on the polymer concentrations.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1993.10a
• /
• pp.482-487
• /
• 1993

A prediction method for determining the welding residual stress by artificial neural network is proposed. A three-dimensional transient thermomechanical analysis has been performed for the CO $_{2}$ Arc Welding using the finite element method. The validity of the above results is demonstrated by experimental elastic stress relief method which is called Holl Drilling Method. The first part of numarical analysis performs a three-dimensional transient heat transfer anslysis, and the second part then uses results of the first part and performs a three-dimensional transient thermo-clasto-plastic analysis to compute transient and residual stresses in the weld. Data from the finite element method were used to train a backpropagation neural network to predict residual stress. Architecturally, the finite element method were used to train a backpropagation voltage and the current, a hidden layer to accommodate failure mechanism mapping, and an output layer for residual stress. The trained network was then applied to the prediction of residual stress in the four specimens. The results of predicted residual stress have been very encouraging.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.30.2-30
• /
• 2001

As the era of the human friendly robot looms, the intelligent autonomous mobile robots have obtained tremendous interests in recent years. The robots may be service robots for serving human or industrial robots for replacing human. For the coexistance with human, the robots must be able to feel and recognize three dimensional space that human live. In this paper, we propose three dimensional environmental modeling method based on a neural network technique called Growing Neural Gas Network. The purpose of this neural network is to generate a graphical structure which reflects the topology of the input space. Through this method, the robots´ surroundings are autonomously segmented ...

• Journal of Welding and Joining
• /
• v.13 no.3
• /
• pp.77-88
• /
• 1995

A prediction method for determining the welding residual stress by artificial neural network is proposed. A three-dimensional transient thermomechanical analysis has been performed for the CO$_{2}$ arc welding using the finite element method. The first part of numerical analysis performs a three-dimensional transient heat transfer analysis, and the second part then uses the results of the first part and performs a three-dimensional transient thermo-elastic-plastic analysis to compute transient and residual stresses in the weld. Data from the finite element method are used to train a backpropagation neural network to predict the residual stress. Architecturally, the fully interconnected network consists of an input layer for the voltage and current, a hidden layer to accommodate the ailure mechanism mapping, and an output layer for the residual stress. The trained network is then applied to the prediction of residual stress in the four specimens. It is concluded that the accuracy of the neural network predicting method is fully comparable with the accuracy achieved by the traditional predicting method.

• International Journal of Korean Welding Society
• /
• v.1 no.2
• /
• pp.51-60
• /
• 2001

A prediction method for determining the welding residual stress by artificial neural network is proposed. A three-dimensional transient thermo-mechanical analysis has been performed for the $CO_2$ arc welding using the finite element method. The first part of numerical analysis performs a three-dimensional transient heat transfer analysis, and the second part then uses the results of the first part and performs a three-dimensional transient thermo-elastic-plastic analysis to compute transient and residual stresses in the weld. Data from the finite element method are used to train a back propagation neural network to predict the residual stress. Architecturally, the fully interconnected network consists of an input layer for the voltage and current, a hidden layer to accommodate the failure mechanism mapping, and an output layer for the residual stress. The trained network is then applied to the prediction of residual stress in the four specimens. It is concluded that the accuracy of the neural network predicting method is fully comparable with the accuracy achieved by the traditional predicting method.

• International Journal of Control, Automation, and Systems
• /
• v.1 no.1
• /
• pp.134-141
• /
• 2003

A major research issue associated with service robots is the creation of an environment recognition system for mobile robot navigation that is robust and efficient on various environment situations. In recent years, intelligent autonomous mobile robots have received much attention as the types of service robots for serving people and industrial robots for replacing human. To help people, robots must be able to sense and recognize three dimensional space where they live or work. In this paper, we propose a three dimensional environmental modeling method based on an edge enhancement technique using a planar fitting method and a neural network technique called "Growing Neural Gas Network." Input data pre-processing provides probabilistic density to the input data of the neural network, and the neural network generates a graphical structure that reflects the topology of the input space. Using these methods, robot's surroundings are autonomously clustered into isolated objects and modeled as polygon patches with the user-selected resolution. Through a series of simulations and experiments, the proposed method is tested to recognize the environments surrounding the robot. From the experimental results, the usefulness and robustness of the proposed method are investigated and discussed in detail.in detail.