Optik
124 (2013) 6020–
6026
Contents
lists
available
at
ScienceDirect
Optik
j
o
ur
nal
hom
epage:
www.elsevier.de/ijleo
Region-based
object
and
background
extraction
via
active
contours
Hui
Wang
a,
b
,
Ting-Zhu
Huang
a,∗
a
School
of
Mathematical
Sciences/Institute
of
Computational
Science,
University
of
Electronic
Science
and
Technology
of
China,
Chengdu,
Sichuan
611731,
China
b
Department
of
Mathematics
and
Computer
Science,
Anshun
University,
Anshun,
Guizhou
561000,
China
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
28
November
2012
Accepted
20
April
2013
Keywords:
Image
segmentation
Active
contour
Level
set
method
Chan–Vese
model
a
b
s
t
r
a
c
t
In
this
paper,
we
propose
a
region-based
model
for
the
object
and
background
extraction
with
application
to
the
image
with
thick
or
complex
boundary.
Based
on
region
information
of
the
image,
we
employ
two
curves
to
extract
the
object
and
background,
respectively,
regardless
of
the
boundary.
The
first
curve
is
used
to
extract
the
object.
Correspondingly,
the
second
curve
is
used
to
extract
the
background.
By
employing
two
level
set
functions
to
represent
the
two
curves,
we
propose
a
new
region-based
energy
functional.
In
the
proposed
model,
a
distance
constraint
term
is
incorporated,
which
effectively
avoid
that
the
two
level
set
functions
too
away
from
each
other
and
keep
their
similar
shapes
well.
Besides,
we
present
a
penalty
term
to
maintain
the
accurate
computation
and
stability
evolution.
Experiment
results
demonstrate
the
desirable
performance
of
the
proposed
model
with
application
to
synthetic
and
real-world
images.
© 2013 Elsevier GmbH. All rights reserved.
1.
Introduction
Image
segmentation
is
a
fundamental
imaging
problem
and
a
large
number
of
different
approaches
to
segmentation
have
been
put
forward
continuously
in
the
past
decades
[1,2].
Since
the
con-
tribution
of
Kass
et
al.
[3],
and
the
introduction
of
level
set
method
[4],
active
contour
models
[3,5–7]
have
been
popular
with
applica-
tion
to
image
segmentation,
which
are
mainly
used
to
extract
the
object
during
a
process
of
minimizing
energy
functional.
Generally
speaking,
active
contour
models
can
be
roughly
categorized
into
edge-based
models
[6,8–10]
and
region-based
models
[7,11–14].
Edge-based
models
employ
local
image
gradient
information
to
attract
the
active
contour
toward
the
object
boundary
and
stop
there.
Among
them,
geodesic
active
contour
(GAC)
model
[6]
is
a
famous
example.
Even
with
a
good
motivation,
this
kind
of
models
are
generally
sensitive
to
initial
conditions
and
noise.
Sometimes
segmenting
a
noisy
image
needs
to
add
a
smoothing
step
prior
seg-
mentation,
bur
doing
this
also
smoothes
image
edges.
Comparing
with
edge-based
models,
region-based
models
incorporate
region
information
to
design
the
fitting
energy.
These
models
are
robust
to
noise
and
specially
able
to
segment
the
object
with
either
fuzzy
or
smooth
edges.
Among
region-based
models,
Mumford–Shah
model
[15]
is
well
known
by
using
the
piecewise
smooth
function
to
approximate
the
image
region.
Whereas,
solving
this
model
is
difficult.
As
its
∗
Corresponding
author.
E-mail
addresses:
wanghui561403@163.com
(H.
Wang),
tingzhuhuang@126.com
(T.-Z.
Huang).
simplified
case,
Chan–Vese
(C–V)
model
[7]
is
proposed
for
two-
phase
image
segmentation.
The
C–V
model
approximates
image
regions
by
two
piecewise
constants,
which
are
the
means
of
inten-
sity
inside
and
outside
the
contour,
respectively.
More
specifically,
it
divide
the
image
into
two
parts.
The
inner
region
is
the
object
and
the
outside
region
is
the
background.
Furthermore,
based
on
the
multiphase
level
set
frame,
Vese
and
Chan
[16]
and
Tsai
et
al.
[12]
independently
presented
two
similar
region-based
models,
which
are
effective
to
segment
the
image
with
complex
regions.
In
[17],
Zhang
et
al.
proposed
a
new
region-based
signed
pressure
force
function,
which
can
efficiently
stop
the
contour
at
weak
or
blurred
edges
with
a
fast
convergence
rate.
In
this
paper,
we
propose
a
new
region-based
active
contour
model,
which
is
used
to
extract
the
object
and
background,
respec-
tively,
regardless
of
the
boundary.
According
to
the
human
visual
perception,
for
the
image
with
thick
or
complex
boundary,
it
is
not
desirable
to
only
divide
the
image
into
two
parts:
object
and
back-
ground.
Generally
speaking,
we
can
not
directly
think
the
boundary
as
the
object
or
background
for
this
kind
of
images.
To
deal
with
this
problem
for
image
segmentation,
we
mainly
consider
to
employ
two
curves
in
this
paper.
Unlike
the
C–V
model,
based
on
the
mini-
mization
of
energy
functional,
the
first
curve
is
driven
to
segment
the
object
and
correspondingly
the
second
curve
aims
to
extract
the
background.
In
other
words,
for
this
kind
of
images,
we
princi-
pally
pay
attention
to
the
object
and
background,
irrespective
of
the
boundary.
More
specifically,
in
the
proposed
model,
we
bring
in
two
level
set
functions,
and
use
them
to
represent
the
two
curves.
Sim-
ilarly
to
[7,18,19],
we
use
the
piecewise
constants
to
approximate
the
regions
of
the
object
and
background,
respectively.
Besides,
a
distance
constraint
term
is
incorporated
into
the
proposed
model
0030-4026/$
–
see
front
matter ©
2013 Elsevier GmbH. All rights reserved.
http://dx.doi.org/10.1016/j.ijleo.2013.04.079
资源评论
码流怪侠
- 粉丝: 2w+
- 资源: 284
最新资源
- (论文+配置方法+源代码+PPT模板)租房宝一站式服务平台(java毕业设计).zip
- java毕业设计丨租赁无忧助手系统(论文+源代码+PPT模板).zip
- (源代码+论文+PPT模板+配置方法)租房宝智能匹配平台【java毕业设计】.zip
- java毕业设计】房源雷达租赁追踪系统(源代码+论文+配置方法+PPT模板).zip
- java毕业设计丨房屋易租管家系统(论文+源代码+PPT模板).zip
- Dism++V10.1.1000.80.zip
- (论文+PPT模板+配置+源代码)租赁无忧一站式管理系统【java毕业设计】.zip
- (源代码+论文+PPT模板+配置方法)房屋租赁智汇系统【java毕业设计】.zip
- java毕业设计】智慧租房助手平台(源代码+论文+PPT模板).zip
- IP地址转换技术及其在网络通信中的应用
- 西安市2005-2024年近20年历史气象数据下载
- (论文+PPT模板+配置+源代码)房源直租信息平台【java毕业设计】.zip
- (源代码+论文+PPT模板+配置方法)租房通全程管理系统【java毕业设计】.zip
- java毕业设计】租赁达人高效管理系统(源代码+论文+PPT模板).zip
- java毕业设计】租房神器智能匹配系统(源代码+论文+PPT模板).zip
- java毕业设计】智慧房屋租赁管家(源代码+论文+PPT模板).zip
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
