1/2005 - 17 | View TOC | « Previous Article | Next Article » |
Adaptive Skin Color Detection in Unconstrained Environments Using 2D Histogram PartitioningRadu Daniel VATAVU, Laurent GRISONI, Samuel DEGRANDE, Christophe CHAILLOU, Stefan Gheorghe PENTIUC |
Extra paper information in |
Click to see author's profile in SCOPUS, IEEE Xplore, Web of Science |
Not available online | Views: 1,388 |
Author keywords
skin color, two dimensional histogram, thresholding, multilayered perceptron
References keywords
References keywords will be displayed on the next page reload.
About this article
Date of Publication: 2005-04-02
Volume 5, Issue 1, Year 2005, On page(s): 101 - xx
ISSN: 1582-7445, e-ISSN: 1844-7600
Digital Object Identifier: Not assigned
Abstract An adaptive approach for skin color detection is presented based on image filtering with dynamically computed thresholds. Two dimensional hue/saturation histograms are analysed and partitioned into rectangular blocks. For each block, simple parameters are computed such as: 1st order moments, standard deviations and simple ratios. A multilayered perceptron is trained using a set of images containing both skin and non-skin blocks. For each frame in a video sequence, each block is assigned a probability of being skin and the parameters of the winning block, if any, are used for image thresholding. |
References | | | Cited By |
On-line references are not available - see the PDF file if available.
Faculty of Electrical Engineering and Computer Science
Stefan cel Mare University of Suceava, Romania
All rights reserved: Advances in Electrical and Computer Engineering is a registered trademark of the Stefan cel Mare University of Suceava. No part of this publication may be reproduced, stored in a retrieval system, photocopied, recorded or archived, without the written permission from the Editor. When authors submit their papers for publication, they agree that the copyright for their article be transferred to the Faculty of Electrical Engineering and Computer Science, Stefan cel Mare University of Suceava, Romania, if and only if the articles are accepted for publication. The copyright covers the exclusive rights to reproduce and distribute the article, including reprints and translations.
Permission for other use: The copyright owner's consent does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific written permission must be obtained from the Editor for such copying. Direct linking to files hosted on this website is strictly prohibited.
Disclaimer: Whilst every effort is made by the publishers and editorial board to see that no inaccurate or misleading data, opinions or statements appear in this journal, they wish to make it clear that all information and opinions formulated in the articles, as well as linguistic accuracy, are the sole responsibility of the author.