crta
Hrvatska znanstvena Sekcija img
bibliografija
3 gif
 Home
 About the project
 FAQ
 Contact
4 gif
Browsing
Basic search
Advanced search
Statistical data
Other bibliographies
Similar projects
 Catalogues and databases

Bibliographic record number: 208920

Journal

Authors: Rimac-Drlje, Snježana; Keller, Alen; Hocenski, Željko
Title: Neural network based detection of defects in texture surfaces
( Neural network based detection of defects in texture surfaces )
Source: Preceedings of IEEE International Symposium on Industrial Electronics ISIE 2005 / Perić, Nedjeljko, Petrović, Ivan, Butković, Željko (ed). - Zagreb : The Institute of Electical and Electronic Engineerings, Inc. , 2005. 1255-1261.
Meeting: IEEE International Symposium on Industrial Electronics ISIE 2005
Location and date: Dubrovnik, Hrvatska, 20-23.06.2005.
Keywords: defect detection; wavelets; neural network; radial basis
( defect detection; wavelets; neural network; radial basis )
Abstract:
In this article we present an algorithm for automatic detection of surface defects on ceramic tiles. This algorithm is based on the probabilistic neural network with radial basis. To improve sensitivity of the detection procedure an image of the tile is divided into segments and one neural network is made for each segment. The discrete wavelet transform (DWT) is used for the feature extraction in every segment. Maximums of the wavelet coefficients as well as the mean value of the approximation coefficients form an input vector for the neural network. Experimental results of the defect detection for different types of tiles and with different parameters of the algorithm show a high sensitivity and applicability of the proposed procedure.
Type of meeting: Predavanje
Type of presentation in a journal: Full-text (1500 words and more)
Type of peer-review: International peer-review
Project / theme: 0165103
Original language: eng
Category: Znanstveni
Research fields:
Computer science



Print version   za tiskati


upomoc
foot_4