By Topic

Neural Network Applications in Electrical Engineering, 2008. NEUREL 2008. 9th Symposium on

Date 25-27 Sept. 2008

Filter Results

Displaying Results 1 - 25 of 89
  • [Front cover]

    Publication Year: 2008 , Page(s): c1
    Save to Project icon | Request Permissions | PDF file iconPDF (7827 KB)  
    Freely Available from IEEE
  • Committees

    Publication Year: 2008 , Page(s): i
    Save to Project icon | Request Permissions | PDF file iconPDF (167 KB)  
    Freely Available from IEEE
  • Foreword

    Publication Year: 2008 , Page(s): ii
    Save to Project icon | Request Permissions | PDF file iconPDF (591 KB) |  | HTML iconHTML  
    Freely Available from IEEE
  • 2008 9th Symposium on Neural Network Applications in Electrical Engineering, Neurel-2008 Conference Proceedings

    Publication Year: 2008 , Page(s): iii
    Save to Project icon | Request Permissions | PDF file iconPDF (83 KB)  
    Freely Available from IEEE
  • Table of contents

    Publication Year: 2008 , Page(s): iv - vi
    Save to Project icon | Request Permissions | PDF file iconPDF (1095 KB)  
    Freely Available from IEEE
  • [Blank page]

    Publication Year: 2008 , Page(s): vii
    Save to Project icon | Request Permissions | PDF file iconPDF (7 KB)  
    Freely Available from IEEE
  • Tutorial lecture 1:

    Publication Year: 2008 , Page(s): viii
    Save to Project icon | Request Permissions | PDF file iconPDF (109 KB)  
    Freely Available from IEEE
  • [Blank page]

    Publication Year: 2008 , Page(s): ix
    Save to Project icon | Request Permissions | PDF file iconPDF (7 KB)  
    Freely Available from IEEE
  • Swarming and flocking: Cooperative collective behavior

    Publication Year: 2008 , Page(s): 1
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (496 KB)  

    The tutorial starts with a presentation of the notion of complexity discussed within the fields of Artificial Intelligence, Neural Networks, Chaos Theory, Self-Organization, Non-linear systems, Emergence and Collective Intelligence. Examples of complex systems include ant-hills, human economies, climate, nervous systems, cells and living things, human beings, as well as modern telecommunication infrastructures, where simple units together behave in complicated ways. Swarm intelligence, which represents an artificial intelligence technique based on the study of collective behavior in self-organized systems, i.e., collective intelligence of groups of simple agents, is also discussed, including the techniques of Ant Colony Optimization and Particle Swarm Optimization. The discussion then encompasses Multi-agent systems, systems that consist of multiple agents or vehicles with several sensors/actuators and the capability to communicate with one another to perform coordinated tasks. The Consensus problem is presented in detail, on the basis of the concepts of Graph theory. Modeling of bird flocking is presented as an example. Decentralized control of autonomous vehicles is taken as another example, with an accent on Formation Control and Sensor Networks. A detailed discussion covers the vehicle and multi-robot formation stability problem, as well as decentralized estimation and control problems in Multi-agent systems based the implementation of a consensus strategy. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • [Blank page]

    Publication Year: 2008
    Save to Project icon | Request Permissions | PDF file iconPDF (7 KB)  
    Freely Available from IEEE
  • Special session 1: COST 292, semantic multimodal analysis of digital media

    Publication Year: 2008 , Page(s): 2a
    Save to Project icon | Request Permissions | PDF file iconPDF (115 KB)  
    Freely Available from IEEE
  • [Blank page]

    Publication Year: 2008 , Page(s): 2b
    Save to Project icon | Request Permissions | PDF file iconPDF (7 KB)  
    Freely Available from IEEE
  • Brief review of COST Action 292 activities

    Publication Year: 2008 , Page(s): 3 - 6
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (4688 KB) |  | HTML iconHTML  

    This paper describes briefly main activities of the COST Action 292 “Semantic multimodal analysis of digital media”. This Action started in 2004 as a logical continuation and extension of previous Action 211. Results from this Action are very promissing and potentially useful for industry in the field of multimedia. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Content based image retrieval: From pixels to semantics

    Publication Year: 2008 , Page(s): 7 - 12
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1612 KB) |  | HTML iconHTML  

    In this paper an evolution of content based image retrieval (CBIR) techniques is presented, together with the most important definitions and problems that have significant importance for the previous development as well as future trends and applications of the discipline. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multimodal semantic characterization of images using MPEG-7 descriptors

    Publication Year: 2008 , Page(s): 13 - 16
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (6161 KB) |  | HTML iconHTML  

    In this paper the edge histogram descriptor, the scalable colour descriptor and the colour layout descriptor defined in the MPEG-7 standard are used for image semantic characterization. A comparative study of the performance and reliability of the image classification based in these descriptors is made. For that, classification methods like neural networks and k-nearest neighbors were used to detect relevant semantic features in images. The descriptors are individually used and combined with different multimodal techniques. A set with 460 images will be used for testing together with a set of 320 training images selected from the TRECVID 2008 development sound and vision database was used. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Enriched access to digital audiovisual content

    Publication Year: 2008 , Page(s): 17 - 20
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (7819 KB) |  | HTML iconHTML  

    This paper presents access engine to digital audio and related content developed under IST FP6 project EASAIER. The main driving force for the project was the lack of qualitative solutions for access to digital sound archives. An innovative remote access system which extends beyond standard content management and retrieval systems, addresses a range of issues identified, such as inconsistent formats of archived materials with related media often in separate collections and related metadata given in non-standard specialist format, incomplete or even erroneous. The system focuses on sound archives, libraries, museums, broadcast archives, and music schools, but the tools may be used by anyone interested in accessing archived material; amateur or professional, regardless of the material involved. The system functionalities; enhanced cross media retrieval, multi-media synchronisation, audio and video processing, analysis and visualisation tools, enable the user to experiment with the materials in exciting new ways. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The use of unlabeled data in image retrieval with relevance feedback

    Publication Year: 2008 , Page(s): 21 - 26
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (6942 KB) |  | HTML iconHTML  

    This paper describes a content-based image retrieval (CBIR) system which makes use of both labeled images, annotated by the user, and unlabeled images available in the database. The system initially retrieves images objectively closest to the query image. The user then subjectively labels retrieved images as relevant or irrelevant. Although such relevance feedback from the user is an effective way of bridging the semantic gap between objective and subjective similarity, it is also very time consuming, requiring huge human effort. Often, the number of labeled images is very small. In an inductive approach the labeled set of images is used for training a CBIR system while the large set of unlabeled images remains unused. In this paper we exploit the transductive support vector machine (SVM) algorithm as a way of taking advantage of unlabeled data in CBIR. Our findings are compared to the results of an inductive SVM. We draw some conclusions as to when the use of unlabeled data might be helpful. The considered systems are tested over images from the Corel 1K dataset. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Image segmentation method based on self-organizing maps and K-means algorithm

    Publication Year: 2008 , Page(s): 27 - 30
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (5148 KB) |  | HTML iconHTML  

    In this paper, a method for color image segmentation based on Kohonenpsilas neural networks and clusterization by using modification of k-means algorithm, is presented. The method consists of three steps. First step includes usage of self-organizing maps for determination of potential candidates for regions centers. Secondly, using maxmin algorithm, number of candidates is reduced to initializing number of centers, which are then used for further analysis. During this process, initial number of regions is formed. For every formed region spatial and intensity centers are determined. Finally, in the third step, iterative procedure of modified k-means algorithm is realized during which the number of regions is minimized. The experimental results verify the usability of described algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Session 1: Algorithms and learning methods

    Publication Year: 2008 , Page(s): 30a
    Save to Project icon | Request Permissions | PDF file iconPDF (85 KB)  
    Freely Available from IEEE
  • [Blank page]

    Publication Year: 2008 , Page(s): 30b
    Save to Project icon | Request Permissions | PDF file iconPDF (7 KB)  
    Freely Available from IEEE
  • Generalized PSO algorithm — an application to Lorenz system identification by means of neural-networks

    Publication Year: 2008 , Page(s): 31 - 35
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3268 KB) |  | HTML iconHTML  

    In this paper a new, generalized PSO (GPSO) algorithm is presented and analyzed, both theoretically and empirically. The new optimizer enables direct control over the properties of the search process. In addition, PSO is addressed in conceptually different manner, revealing further aspects of the algorithm behavior. GPSO is applied for training radial basis function neural network (RBF-NN) to identify dynamics of a nonlinear system. The target system is chosen to be of Lorenz type, known for its complex, chaotic behavior. Results presented in this paper clearly demonstrate effectiveness of the proposed algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • [Blank page]

    Publication Year: 2008 , Page(s): 36
    Save to Project icon | Request Permissions | PDF file iconPDF (0 KB)  
    Freely Available from IEEE
  • Sequential training of Support Vector Machine

    Publication Year: 2008 , Page(s): 37 - 42
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (4427 KB) |  | HTML iconHTML  

    In this paper we present efficient implementation of an algorithm for sequential training of support vector machine. Algorithm is obtained by maintaining Karush-Kuhn-Tucker optimality conditions while learning from new example. We have tested the performance of our implementation on widely recognized classification benchmark tests. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Named entity recognition and classification using context Hidden Markov Model

    Publication Year: 2008 , Page(s): 43 - 46
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1177 KB) |  | HTML iconHTML  

    Named entity (NE) recognition is a core technology for understanding low level semantics of texts. In this paper we report our preliminary results for Named Entity Recognition on MUC 7 corpus by combining the supervised machine learning system in the form of probabilistic generative Hidden Markov Model (HMM) for named entity classes PERSON, ORGANIZATION and LOCATION, and grammar based component for DATE, TIME, MONEY and PERCENT. We have implemented two variations of the basic Hidden Markov Model, where the second one is our version of HMM which uses the context of surrounding words to determine the NE class of the current word, leading to more accurate and faster NE recognition. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Kolmogorov complexity of spherical vector quantizers

    Publication Year: 2008 , Page(s): 47 - 52
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1392 KB) |  | HTML iconHTML  

    In this paper, we investigate memory complexity of spherical vector quantizer from Kolmogorovpsilas perspective. The method for expressing the quantizer as binary string is proposed and minimal description length of the string is considered as Kolmogorov complexity of the quantizer. The Kolmogorov complexity is compared to memory requirements of two main algorithms for spherical vector quantizer design: uniform spherical quantizer and generalized Lloyd-Maxpsilas algorithm. It is proven that first of them has the minimal memory requirements needed for spherical quantizer realization, while the other upper bounds the theoretical minimal description length of the quantizer. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.