Publication detail
COMPARISON OF NEURAL NETWORKS FOR SKIN SEGMENTATION IN COLOR IMAGES
HYNČICA, T. JIRSÍK, V.
Original Title
COMPARISON OF NEURAL NETWORKS FOR SKIN SEGMENTATION IN COLOR IMAGES
English Title
COMPARISON OF NEURAL NETWORKS FOR SKIN SEGMENTATION IN COLOR IMAGES
Type
conference paper
Language
en
Original Abstract
This paper describes comparison of results of skin pixels segmentation in color images. Two types of artificial neural networks and three color spaces were used in this work. The neural networks selected in this study, Reduced Coulomb Energy (RCE) and cosine reformulated radial basis function (RBF) neural networks, have similar structure and both divide the space using multi-dimensional spheres. Since this work focuses on segmentation based on color, three color spaces frequently used in image processing (HSL, YCbCr and L*a*b*) were tested for their suitability to represent skin colors.
English abstract
This paper describes comparison of results of skin pixels segmentation in color images. Two types of artificial neural networks and three color spaces were used in this work. The neural networks selected in this study, Reduced Coulomb Energy (RCE) and cosine reformulated radial basis function (RBF) neural networks, have similar structure and both divide the space using multi-dimensional spheres. Since this work focuses on segmentation based on color, three color spaces frequently used in image processing (HSL, YCbCr and L*a*b*) were tested for their suitability to represent skin colors.
Keywords
RBF neural network, RCE neural network, color spaces, skin segmentation
RIV year
2013
Released
26.06.2013
Location
Koperníkova 4362/21, Brno, Czech Republic
ISBN
978-80-214-4755-4
Book
MENDEL 2013
Pages from
203
Pages to
208
Pages count
6
Documents
BibTex
@inproceedings{BUT101180,
author="Tomáš {Hynčica} and Václav {Jirsík}",
title="COMPARISON OF NEURAL NETWORKS FOR SKIN SEGMENTATION IN COLOR IMAGES",
annote="This paper describes comparison of results of skin pixels segmentation in color images. Two types of artificial neural networks and three color spaces were used in this work. The neural networks selected in this study, Reduced Coulomb Energy (RCE) and cosine reformulated radial basis function (RBF) neural networks, have similar structure and both divide the space using multi-dimensional spheres. Since this work focuses on segmentation based on color, three color spaces frequently used in image processing (HSL, YCbCr and L*a*b*) were tested for their suitability to represent skin colors.",
booktitle="MENDEL 2013",
chapter="101180",
howpublished="print",
year="2013",
month="june",
pages="203--208",
type="conference paper"
}