Most edge detection algorithms include three main stages: smoothing, differentiation, and labeling. In this paper, we evaluate the performance of algorithms in which competitive learning is applied first to enhance ed...
详细信息
ISBN:
(纸本)1604238216
Most edge detection algorithms include three main stages: smoothing, differentiation, and labeling. In this paper, we evaluate the performance of algorithms in which competitive learning is applied first to enhance edges, followed by an edge detector to locate the edges. In this way, more detailed and relatively more unbroken edges can be found as compared to the results when an edge detector is applied alone. The algorithms compared are K-Means, SOM and SOGR for clustering, and Canny and GED for edge detection. Perceptionally, best results were obtained with the GED-SOGR algorithm. The SOGR is also considerably simpler and faster than the SOM algorithm.
In this paper, we discuss edge detection by first using a clustering algorithm followed by a known edge detection filter such as Canny or Generalized Edge Detector (GED). We developed a new clustering method called Se...
详细信息
Most edge detection algorithms include three main stages: smoothing, differentiation, and labeling. In this paper, we evaluate the performance of algorithms in which competitive learning is applied first to enhance ed...
详细信息
Most edge detection algorithms include three main stages: smoothing, differentiation, and labeling. In this paper, we evaluate the performance of algorithms in which competitive learning is applied first to enhance edges, followed by an edge detector to locate the edges. In this way, more detailed and relatively more unbroken edges can be found as compared to the results when an edge detector is applied alone. The algorithms compared are K-Means, SOM and SOGR for clustering, and Canny and GED for edge detection. Perceptionally, best results were obtained with the GED-SOGR algorithm. The SOGR is also considerably simpler and faster than the SOM algorithm.
Edge detection is an important topic in image processing and a main tool in pattern recognition and image segmentation. Many edge detection techniques are available in the literature. 'A number of recent edge dete...
详细信息
Edge detection is an important topic in image processing and a main tool in pattern recognition and image segmentation. Many edge detection techniques are available in the literature. 'A number of recent edge detectors are multiscale and include three main processing steps: smoothing, differentiation and labeling' (Ziau and Tabbone, 1997). This paper, presents a proposed method which is suitable for edge detection in images. This method is based on the use of the clustering algorithms (Self-Organizing Map (SOM), K-Means) and a gray scale edge detector (Canny, Generalized Edge Detector (GED)). It is shown that using the grayscale edge detectors may miss some parts of the edges which can be found using the proposed method.
作者:
Okano, KHigashino, TTaniguchi, KMemberFaculty of Engineering Science
Osaka University Toyonaka Japan 560 Teruo Hipshino received his B.E.
M.E.and Ph.D. degrees in Information and Computer Sciences from Osaka University Osaka Japan in 1R9 1981 and 1984 respectively. He joined the Faculty of Osaka University in 1984. Since 1991 he has been an Associate Professor in the Department of Information and Computer Sciences at Osaka University. In 1990 he was a Visiting Researcher of Dept. I.R.O. at University of Montreal Canada. His current research interests include design and analysis of distributed systems specification and verification of communication protocols and formal approach of program design. He is a member of IEEE-CS ACMIEICE of Japanand IPS of Japan.
In this paper, new methods for providing the correctness of refinement among abstract sequential machine style programs are described. The programs are described in algebraic language ASL using some useful notions, i....
详细信息
In this paper, new methods for providing the correctness of refinement among abstract sequential machine style programs are described. The programs are described in algebraic language ASL using some useful notions, i.e., the extended projection and the valid reachability condition for each transition function. These notions allow a designer to refine a given text (program or specification) to a concrete text more freely than the text that does not use such notions. These notions can also enhance the expressive power of the text. On the other hand, these advantages would have lost half their values, if useful methods to prove the correctness of refinement among the texts using them are not found. Thus new methods for proving the correctness of the texts are proposed, and they do not require much proof loads. First, the correctness of refinement among the ASL texts with the extended projection is defined. Second, a method is proposed for proving the correctness of refinement among such texts. Also proposed is a definition of the correctness of refinement among the texts with the valid reachability conditions and a method to prove their correctness. These methods do not require more proof loads than the usual methods for the texts that do not use such notions in the abstract sequential machine style. Therefore, these methods are useful.
作者:
Zhuang, XPKoshiba, MTsuji, YFaculty of Engineering
Hokkaido University Sapporo Japan 060 Graduated from the Department of Communication Engineering
Guellin Electronic Institute China in 1986. She received her M.S. degree from Hokkaido University in 1991 and was admitted to the doctoral program. She has been engaged in research on the optimum design of magnetooptic waveguides for nonreciprocal optical devices. Graduated in 1971 from the Department of Electronic Engineering
Hokkaido University where he received his M.S. degree in 1973 and his Dr. of Eng. degree in 1976. In 1976 he became a Lecturer in the Department of Electronic Engineering Kitami Institute of Technology where he was promoted to Associate Professor in 1977. In 1979 he became an Associate Professor at Hokkaido University where he was promoted to Professor in 1987. He has been engaged in research on optical and wave electronics. In 1987 he received a Best Paper Award. He is the author ofFundamentals of Finite Element Method of Optics and Waves(Morigkata Publ.)Optical Waveguide Analysis(Asakura Publ.)Optical Waveguide Analysis(McGraw-Hill Book Co.) andOptical Waveguide Theory by the Finite Element Method(KTK Scientific Publishing/Kulwer Academic Publishers). He is also co-author of one book and has written chapters in six other books. He is a member of the Institute of Television Engineers of Japan the Institute of Electrical Engineers of Japan the Japan Society for Simulation Technology the Japan Society for Computational Methods in Engineering and the Japan Society of Applied Magnetics and Mechanics. He is also an IEEE Senior Member. Graduated in 1991 from me Department of Electronic Engineering
Hokkaido University where he received his M.S. degree in 1993. He is currently in the doctoral program. He has been engaged in research on quantum wave phenomena and computer-aided design of optical and quantum effect devices.
To evaluate nonreciprocal mode conversion characteristics in a magnetooptic channel waveguide, a finite element analysis based on the scalar wave approximation is formulated for the first time. A simple iterative calc...
详细信息
To evaluate nonreciprocal mode conversion characteristics in a magnetooptic channel waveguide, a finite element analysis based on the scalar wave approximation is formulated for the first time. A simple iterative calculation method is conceived for solution of the nonlinear eigenvalue equations finally obtained in which the two polarizations are coupled. Specifically, the maximum isolation ratio in the mode-conversion-type magnetooptic rib guide was evaluated and the validity of the method is confirmed by comparison with published experimental results.
作者:
IKUNO, HGONDO, MNISHIMOTO, MFaculty of Engineering
Kumamoto University Kumamoto Japan 860 Himyoshi Ikmo received his B.E.
M.E. and Ph.D. from Kyushu University in 1962 1964 and 1979 respectively. From 1967 to 1970 he was a Research Associate in Computer Science and Communication Engineering at Kyushu University. He joined the Department of Electrical Engineering and Computer Science of Kumamoto University in 1970 and is currently a Professor there. During 1983 to 1984 he was a Visiting Professor at the Polytechnic University Brooklyn NY. His research interests are in the areas of scattering diffraction and wave propagation in the theory of electromagnetics. Dr. Ikuno is a member of IEEE and IEE of Japan. Michim Gondo graduated from National Defense Academy
Yokosuka in 1985 and received his M.E. from Kumamoto University in 1990. He is currently in the doctoral program at Kumamoto University. His research interest is in three-dimensional electromagnetic scattering. Masahiko Nkhimoto received his B.S. in Electronic Engineering from Kumamoto University in 1982 and an M.S. and Ph.D. from Kyushu University in 1984 and 1987
respectively. Since 1987 he has been a Research Associate in the Department of Electrical Engineering and Computer Science Kurnamoto University. His research interests are in the areas of scattering and diffraction of electromagnetic waves.
This paper shows that the surface current density can be calculated by using approximate total tangent magnetic vector fields which are expressed by finite number of linear combination of tangent vectors of spherical ...
详细信息
This paper shows that the surface current density can be calculated by using approximate total tangent magnetic vector fields which are expressed by finite number of linear combination of tangent vectors of spherical wave functions. Sampling points are determined by considering the property of associated Legendre functions and trigonometrical functions. By setting a number of sampling points to twice the largest mode number, a computer-aided algorithm is obtained. The accuracy of the solutions is checked and the polarization dependence of the surface current densities is clarified.
作者:
Maruyama, SatoshiIwanami, YasunoriIkeda, TetsuoFaculty of Engineering
Nagoya Institute of Technology Nagoya Japan 486 Satoshi Maruyama graduated 1983 Dept. Electrical Eng.
Fac. Eng. Nagoya Inst. Technology. Completed Master's program 1985 Grad. School and affiliated with Fujitsu Co. Engaged in researches in distance measurement by spread spectrum technique. Tetsuo Ikeda graduated 1961 Dept. Comm. Eng.
Fac. Eng. Tohoku University. Completed doctoral program 1966 Grad. School. Assistant 1966 and Assoc. Prof. 1968 Fac. Eng. Tohoku University. Assoc. Prof. 1973 and Prof. 1977 Nagoya Inst. Technology. Engaged in researches in network theory computer networks communication system and electromagnetic environment. Doctor of Eng. Member I.E.E.E. Japan TV Soc. Jap. Inf. Proc. Soc. Jap. Soc. Inf. Comm. and I.E.E.E.
The spread spectrum (SS) communication is a communication system which has excellent characteristics against noise and is suited to the code-division multiplex. This paper describes a method which reduces the tracking...
详细信息
The spread spectrum (SS) communication is a communication system which has excellent characteristics against noise and is suited to the code-division multiplex. This paper describes a method which reduces the tracking phase error in DLL due to multipath waves in SS communication system. The method is based on the model with a single multipath wave.
暂无评论