• Journal of the Korea Society of Computer and Information
• /
• v.17 no.12
• /
• pp.83-93
• /
• 2012

This paper proposes an image segmentation framework that modifies the objective function of Fuzzy C-Means (FCM) to improve the performance and computational efficiency of the conventional FCM-based image segmentation. The proposed image segmentation framework includes a locally weighted fuzzy c-means (LWFCM) algorithm that takes into account the influence of neighboring pixels on the center pixel by assigning weights to the neighbors. Distance between a center pixel and a neighboring pixels are calculated within a window and these are basis for determining weights to indicate the importance of the memberships as well as to improve the clustering performance. We analyzed the segmentation performance of the proposed method by utilizing four eminent cluster validity functions such as partition coefficient ($V_{pc}$), partition entropy ($V_{pe}$), Xie-Bdni function ($V_{xb}$) and Fukuyama-Sugeno function ($V_{fs}$). Experimental results show that the proposed LWFCM outperforms other FCM algorithms (FCM, modified FCM, and spatial FCM, FCM with locally weighted information, fast generation FCM) in the cluster validity functions as well as both compactness and separation.

• Journal of Korea Society of Industrial Information Systems
• /
• v.25 no.5
• /
• pp.83-90
• /
• 2020

In this paper, a web-based seed phenotype visualization support system is proposed to extract and visualize data such as the surface color, length, area, perimeter and compactness of seed, which is phenotype information from the image of soybean/rice seeds. This system systematically stores data extracted from seeds in databases, and provides a web-based user interface that facilitates the analysis of data by researchers using data tables and charts. Conventional seed characteristic studies have been manually measured by humans, but the system developed in this paper allows researchers to simply upload seed images for analysis and obtain seed's numerical data after image processing. It is expected that the proposed system will be able to obtain time efficiency and remove spatial restriction, if it is used in seed characterization research, and it will be easy to analyze through systematic management of research results and visualization of the phenotype characteristics.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.61 no.2
• /
• pp.312-320
• /
• 2012

In this study, we introduce a new design methodology of a granular-oriented self-organizing polynomial neural networks (GoSOPNNs) that is based on multi-layer perceptron with Context-based Polynomial Neurons (CPNs) or Polynomial Neurons (PNs). In contrast to the typical architectures encountered in polynomial neural networks (PNN), our main objective is to develop a methodological design strategy of GoSOPNNs as follows : (a) The 1st layer of the proposed network consists of Context-based Polynomial Neuron (CPN). In here, CPN is fully reflective of the structure encountered in numeric data which are granulated with the aid of Context-based Fuzzy C-Means (C-FCM) clustering method. The context-based clustering supporting the design of information granules is completed in the space of the input data while the build of the clusters is guided by a collection of some predefined fuzzy sets (so-called contexts) defined in the output space. (b) The proposed design procedure being applied at each layer of GoSOPNN leads to the selection of preferred nodes of the network (CPNs or PNs) whose local characteristics (such as the number of contexts, the number of clusters, a collection of the specific subset of input variables, and the order of the polynomial) can be easily adjusted. These options contribute to the flexibility as well as simplicity and compactness of the resulting architecture of the network. For the evaluation of performance of the proposed GoSOPNN network, we describe a detailed characteristic of the proposed model using a well-known learning machine data(Automobile Miles Per Gallon Data, Boston Housing Data, Medical Image System Data).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39C no.4
• /
• pp.380-386
• /
• 2014

In this paper, a higher-order mode coupler using coaxial-structure waveguide for Ku-band monopulse satellite tracking is proposed. The proposed higher-order mode coupler is built in a coaxial structure for compactness and weight reduction, making it suitable for mobile tracking systems. The inner circular waveguide of coaxial-structure is used to extract the fundamental mode signal and the high-order mode signal is extracted from the four slots of the outer circular waveguide and then transmitted via given stepped rectangular waveguide structure. The simulated results show that proposed higher-order mode coupler covers 250 MHz(12.75 GHz ~ 13.00 GHz) bandwidth with return loss and insertion loss characteristics. The antenna patterns of fundamental mode and higher-order mode applicable to monopulse tracking are generated successfully. Designed higher-order mode coupler using coaxial waveguide structure for Ku-band is expected to be used for high precision monopulse satellite tracking systems.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.49 no.12
• /
• pp.124-130
• /
• 2012

In this paper, we propose a feature point selection method for MPEG CDVS CE-7 which is processing on International Standard task. Among a large number of extracted feature points, more important feature points which is used in image matching should be selected for the compactness of image descriptor. The proposed method is that remove the feature point in the extraction phase which is filtered by nearest neighbor distance ratio matching in the matching phase. We can avoid the waste of the feature point and employ additional feature points by the proposed method. The experimental results show that our proposed method can obtain true positive rate improvement about 2.3% in pair-wise matching test compared with Test Model.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.4
• /
• pp.485-492
• /
• 2006

This paper deals with implementation of network for distributed control system of a humanoid robot ISHURO(Inha Semyung Humanoid Robot). A humanoid robot needs much degree of freedom structurally and much data for having flexible movement. To realize such a humanoid robot, distributed control method is preferred to the centralized one since it gives a compactness, modularity and flexibility for the controllers. For organizing distributed control system of a humanoid robot, a control processor on a board is needed to individually control the joint motor and communication technology between the processors is required to transmit its information within control time. The processor is DSP-based processor and includes CAN network on a chip. It shares the computational load such as monitoring the sensor information and controlling the actuator between each of modules. In this paper, the communication architecture is suggested and its message protocol are discussed including message structure, time consumption for transmission, and controller structure at the view of distributed control for a humanoid robot. All of the sequence are simulated with Matlab and then verified with real walking experiment by ISHURO.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.29 no.4
• /
• pp.266-269
• /
• 2018

A compact Q-band waveguide-to-microstrip transition for UAV(Unmanned Aerial Vehicle) radiometer applications is presented. The key features of this transition are simplicity, compactness, easy matching, and lower sensitivity to the dimensions and fabrication tolerances. The simple E-plane patch-type design is insensitive to the backshort cavity enclosure and misalignment between the waveguide and microstrip substrate. The primary parameters are optimized using a three-dimensional(3D) electromagnetic simulator(ANSYS HFSS). It exhibited better than 20-dB return loss at mid-band frequencies with less than 1-dB insertion loss for the back-to-back transition, and a return loss better than 15 dB over the frequency range of 36 GHz to 42 GHz.

• Korean Journal of Materials Research
• /
• v.28 no.4
• /
• pp.201-207
• /
• 2018

Infrared(IR) heating has many advantages, such as energy efficiency, reduced heating time, cleanliness, equipment compactness, high drying rate and easy automation. These features of IR heating provide widely industrial applications, such as surface heat treatment in semiconductor fabrication, thermoforming of polymers, drying and disinfection of food products, heating to metal forging, and drying of wet materials. In this study, the characteristics of a protected gold mirror were examined by spectrophotometer and the lifetime of the coating layers were evaluated by a cross-cutting method and salt spray test. The effects of manufacturing conditions on the protected gold mirror were seen and remedies for these effects were noted in order to improve the properties of the protected gold mirror in the drying process. The reflectance and lifetime of the protected gold mirror was influenced by manufacturing conditions, such as surface roughness and forming conditions of the anti-oxide layer, the adhesion layer, the reflecting layer and the protection layer. The results of this study showed that the protected gold mirror manufactured using a buffing method for pre-treatment resulted in the most effective reflectance. In addition, $Al_2O_3$ coating on an Al substrate as an anti-oxide layer was more effective than the anodizing process in the test of reflectance. Furthermore, the protected gold mirror manufactured by layers forming of various materials resulted in the most effective reflectance and lifetime when coated with $Al_2O_3$ as the anti-oxide layer, coated Cr as the adhesion layer, and coated $MgF_2$ as the protection layer.

• Proceedings of the IEEK Conference
• /
• 2001.10a
• /
• pp.198-201
• /
• 2001

Porous structures of aluminum foam have been studied by using return aluminum scrap. The apparent foam shape, foam height, density, pore size and their distributions in various section areas of the experimental samples have been investigated. The sample have been cast into metallic mold, using aluminum foam prepared from a precursor based on pure Al ingot and return aluminum scrap mixed with various amounts of 1-2wt% increasing viscosity and foam agent materials. The process provides for flexibility in design of foam structures via relatively easy control over the amount of hydrogen evolution and the drainage processes which occur during foam formation. This is facilitated by manipulating parameters such as the foaming agent, thermal histories during solidification and mix melt viscosities. A metal for producing the foamed are decomposing a foaming agent in a molten metal such that there is an initial and a subsequent expansion due to foaming agent. It has been found that the Al porous foaming with variation amount of 1∼2wt% foam agent and at 2min holding time, which melting temperature has appeared homogeneous pore size at 650∼700$^{\circ}C$. The compression strength were 10-13 kg/min at 125ppi, and increased by higher pore density. The acoustical performance of the panel made with the foamed aluminum is considerably improved; its absorption coefficient shows NRC 0.6-0.8. It has been found that the Al foam is very preferable for the compactness of the thermal system.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.4
• /
• pp.451-455
• /
• 2008

In this paper, the k-nearest templates method is proposed to combine multiple classifiers effectively. First, the method decomposes training samples of each class into several subclasses based on the outputs of classifiers to represent a class as multiple models, and estimates a localized template by averaging the outputs for each subclass. The distances between a test sample and templates are then calculated. Lastly, the test sample is assigned to the class that is most frequently represented among the k most similar templates. In this paper, C-means clustering algorithm is used as the decomposition method, and k is automatically chosen according to the intra-class compactness and inter-class separation of a given data set. Since the proposed method uses multiple models per class and refers to k models rather than matches with the most similar one, it could obtain stable and high accuracy. In this paper, experiments on UCI and ELENA database showed that the proposed method performed better than conventional fusion methods.