Region competition: Unifying snakes, region growing, and Bayes/MDL
for multiband image segmentation
S. Zhu, and A. Yuille. IEEE Transactions On Pattern Analysis And Machine Intelligence, 18 (9):
884-900(September 1996)
Abstract
We present a novel statistical and variational approach to image
segmentation based on a new algorithm named region competition. This
algorithm is derived by minimizing a generalized Bayes/MDL criterion
using the variational principle. The algorithm is guaranteed to converge
to a local minimum and combines aspects of snakes/balloons and region
growing. Indeed the classic snakes/balloons and region growing algorithms
can be directly derived from our approach. We provide theoretical
analysis of region competition including accuracy of boundary location,
criteria for initial conditions, and the relationship to edge detection
using filters. It is straightforward to generalize the algorithm
to multiband segmentation and we demonstrate it on gray level images,
color images and texture images. The novel color model allows us
to eliminate intensity gradients and shadows, thereby obtaining segmentation
based on the albedos of objects. It also helps detect highlight regions.
%0 Journal Article
%1 Zhu1996
%A Zhu, SC
%A Yuille, A
%C 10662 LOS VAQUEROS CIRCLE, PO BOX 3014, LOS ALAMITOS, CA 90720-1264
%D 1996
%I IEEE COMPUTER SOC
%J IEEE Transactions On Pattern Analysis And Machine Intelligence
%K Bayes color description growing; length; minimum model} principle; region segmentation; snakes; statistics; uncertainty {image
%N 9
%P 884-900
%T Region competition: Unifying snakes, region growing, and Bayes/MDL
for multiband image segmentation
%V 18
%X We present a novel statistical and variational approach to image
segmentation based on a new algorithm named region competition. This
algorithm is derived by minimizing a generalized Bayes/MDL criterion
using the variational principle. The algorithm is guaranteed to converge
to a local minimum and combines aspects of snakes/balloons and region
growing. Indeed the classic snakes/balloons and region growing algorithms
can be directly derived from our approach. We provide theoretical
analysis of region competition including accuracy of boundary location,
criteria for initial conditions, and the relationship to edge detection
using filters. It is straightforward to generalize the algorithm
to multiband segmentation and we demonstrate it on gray level images,
color images and texture images. The novel color model allows us
to eliminate intensity gradients and shadows, thereby obtaining segmentation
based on the albedos of objects. It also helps detect highlight regions.
@article{Zhu1996,
abstract = {{We present a novel statistical and variational approach to image
segmentation based on a new algorithm named region competition. This
algorithm is derived by minimizing a generalized Bayes/MDL criterion
using the variational principle. The algorithm is guaranteed to converge
to a local minimum and combines aspects of snakes/balloons and region
growing. Indeed the classic snakes/balloons and region growing algorithms
can be directly derived from our approach. We provide theoretical
analysis of region competition including accuracy of boundary location,
criteria for initial conditions, and the relationship to edge detection
using filters. It is straightforward to generalize the algorithm
to multiband segmentation and we demonstrate it on gray level images,
color images and texture images. The novel color model allows us
to eliminate intensity gradients and shadows, thereby obtaining segmentation
based on the albedos of objects. It also helps detect highlight regions.}},
added-at = {2011-03-11T12:21:24.000+0100},
address = {{10662 LOS VAQUEROS CIRCLE, PO BOX 3014, LOS ALAMITOS, CA 90720-1264}},
affiliation = {{Zhu, SC (Reprint Author), HARVARD UNIV,DIV APPL SCI,ROBOT LAB,CAMBRIDGE,MA
02138. SMITH KETTLEWELL EYE RES INST,SAN FRANCISCO,CA 94115.}},
author = {Zhu, SC and Yuille, A},
biburl = {https://www.bibsonomy.org/bibtex/2659761c7f8e1155ed3934a6311c3e25e/jmaiora},
doc-delivery-number = {{VK799}},
interhash = {3b0b1d9c6f5b0fa9bc91501bfcdc8780},
intrahash = {659761c7f8e1155ed3934a6311c3e25e},
issn = {{0162-8828}},
journal = {{IEEE Transactions On Pattern Analysis And Machine Intelligence}},
journal-iso = {{IEEE Trans. Pattern Anal. Mach. Intell.}},
keywords = {Bayes color description growing; length; minimum model} principle; region segmentation; snakes; statistics; uncertainty {image},
keywords-plus = {{MODELS; DISTRIBUTIONS; EXTRACTION; ALGORITHM}},
language = {{English}},
month = {{SEP}},
number = {{9}},
number-of-cited-references = {{41}},
owner = {Josu},
pages = {{884-900}},
publisher = {{IEEE COMPUTER SOC}},
subject-category = {{Computer Science, Artificial Intelligence; Engineering, Electrical
\& Electronic}},
times-cited = {{672}},
timestamp = {2011-03-11T12:21:28.000+0100},
title = {{Region competition: Unifying snakes, region growing, and Bayes/MDL
for multiband image segmentation}},
type = {{Article}},
unique-id = {{ISI:A1996VK79900003}},
volume = {{18}},
year = {{1996}}
}