Paper sigradi2010_324:
Segmentación de imágenes digitales mediante técnicas de clustering complementadas con técnicas de crecimiento de regiones

Cumulative Index of Computer Aided Architectural Design
in Spanish

Works

Welcome guest _anon_410478

id sigradi2010_324
authors Toranzo, Lorca Gustavo; Pereira Barzaga Osvaldo; Arzola Ruiz José
year 2010
title Segmentación de imágenes digitales mediante técnicas de clustering complementadas con técnicas de crecimiento de regiones
source SIGraDi 2010_Proceedings of the 14th Congress of the Iberoamerican Society of Digital Graphics, pp. Bogotá, Colombia, November 17-19, 2010, pp. 324-327
summary This paper proposes a solution to the problem of digital image segmentation based on clustering techniques. K - means and fuzzy k - means are clustering algorithms detailed within this study. We propose novel modifications to k - means and fuzzy k - means algorithms to vary objective functions and to use color difference as a distance measure. These new algorithms are guaranteed to obtain the same results as the originals while considerably reducing the execution time. Pixel growing is used to obtain connected structures in the results of clustering algorithms.
keywords clustering; fuzzy k - means; image segmentation; k - means
series SIGRADI
email gtoranzo@uci.cu
content file.pdf (159,130 bytes)
discussion No discussions. Post discussion ...
ratings
last changed 2010/12/11 15:25
 
© CUMINCAD 2004