Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
2
2019 FBSD
Manage
Activity
Members
Labels
Plan
Issues
0
Issue boards
Milestones
Wiki
Code
Merge requests
0
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package Registry
Model registry
Operate
Environments
Terraform modules
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
Kerautret Bertrand
2019 FBSD
Commits
666ce8d2
Commit
666ce8d2
authored
6 years ago
by
even
Browse files
Options
Downloads
Patches
Plain Diff
Article: notions section enhanced
parent
c5a10112
No related branches found
Branches containing commit
No related tags found
No related merge requests found
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
Article/abstract.tex
+3
-3
3 additions, 3 deletions
Article/abstract.tex
Article/notions.tex
+35
-27
35 additions, 27 deletions
Article/notions.tex
with
38 additions
and
30 deletions
Article/abstract.tex
+
3
−
3
View file @
666ce8d2
This paper introduces a new straight edge detector in gray-level images
based on blurred segments, digital objects able to imbed quality measurements
on the extracted features. This study completes previous works with a better
estimation of the blurred segment width and orientation, with t
he help of two
main
improvements
: adaptive directional scans and the control of the
assigned width to the
recogni
tion algorithm.
estimation of the blurred segment width and orientation, with t
wo main
improvements: adaptive directional scans and the control of the
assigned width to the
detec
tion algorithm.
This diff is collapsed.
Click to expand it.
Article/notions.tex
+
35
−
27
View file @
666ce8d2
...
...
@@ -20,8 +20,8 @@ $h_P(\mathcal{L}) = c - ax - by$ its {\it shift} to point $P(x,y)$,
$
h
_
0
(
\mathcal
{
L
}
)
=
c
$
its
{
\it
shift
}
to origin, and
$
p
(
\mathcal
{
L
}
)
=
max
(
|a|,|b|
)
$
its period (i.e. the length of its
periodic pattern).
When
$
\nu
=
p
$
, then
$
\mathcal
{
L
}$
is the narrowest 8-connected
line
and is called a naive line.
When
$
\nu
=
p
(
\mathcal
{
L
}
)
$
, then
$
\mathcal
{
L
}$
is the narrowest 8-connected
line
and is called a naive line.
\begin{definition}
A blurred segment
$
\mathcal
{
B
}$
of assigned width
$
\varepsilon
$
is a set
...
...
@@ -98,13 +98,14 @@ the scan strip is defined by :
\mathcal
{
D
}
(A,B) =
\mathcal
{
L
}
(
\delta
_
x,
\delta
_
y, min (c1,c2), 1 + |c
_
1-c
_
2|)
\end{equation}
\noindent
where
$
\delta
_
x
=
x
_
B
-
x
_
A,
\delta
_
y
=
y
_
B
-
y
_
A,
c
_
1
=
a
\cdot
x
_
A
+
b
\cdot
y
_
A
$
and
$
c
_
2
=
a
\cdot
x
_
B
+
b
\cdot
y
_
B
$
.
where
$
\delta
_
x
=
x
_
B
-
x
_
A
$
,
$
\delta
_
y
=
y
_
B
-
y
_
A
$
,
$
c
_
1
=
\delta
_
x
\cdot
x
_
A
+
\delta
_
y
\cdot
y
_
A
$
and
$
c
_
2
=
\delta
_
x
\cdot
x
_
B
+
\delta
_
y
\cdot
y
_
B
$
.
The scan line
$
\mathcal
{
N
}_
i
$
is then defined by :
\begin{equation}
\mathcal
{
N
}_
i(A,B) =
\mathcal
{
L
}
(
\delta
_
y, -
\delta
_
x,
\delta
_
y
\cdot
x
_
A -
\delta
_
X
\cdot
y
_
A + i
\cdot
\nu
_{
AB
}
,
\nu
_{
AB
}
)
\delta
_
y
\cdot
x
_
A -
\delta
_
x
\cdot
y
_
A + i
\cdot
\nu
_{
AB
}
,
\nu
_{
AB
}
)
\end{equation}
where
$
\nu
_{
AB
}
=
max
(
|
\delta
_
x|, |
\delta
_
y|
)
$
...
...
@@ -129,16 +130,17 @@ and the scan line $\mathcal{N}_i(C,\vec{D},w)$ :
\subsection
{
Adaptive directional scan
}
The blurred segment is searched within a directional scan w
hich
position
and orientation a
re given by the user, or defined in arbitrary direction in
unsupervised mode.
Most of the time
s
, the detection stops where the segment escapes sideways
from
s
the scan strip (
\RefFig
{
fig:escape
}
).
Therefore a
second search is
run agai
n using a
second
directional scan aligned
The blurred segment is searched within a directional scan w
ith a
position
and
an
orientation a
pproximately provided by the user, or blindly defined
in
unsupervised mode.
Most of the time, the detection stops where the segment escapes sideways
from the scan strip (
\RefFig
{
fig:escape
}
).
A
second search is
then ru
n using a
nother
directional scan aligned
on the detected segment.
But only a quantized estimation of this blurred segment direction is given,
and the longer the real segment, the higher the probability to fail again
on a blurred segment escape from the directional scan.
However, even in the case of an exact detection, the estimated orientation
of the segment is subject to the numerization rounding,
and the longer the real segment to detect, the higher the probability to
fail again on a blurred segment escape from the directional scan.
\begin{figure}
[h]
\center
...
...
@@ -174,21 +176,23 @@ on a blurred segment escape from the directional scan.
% \label{fig:cone}
%\end{figure}
In the former work, an additional refinement step was run,
but doing so, the problem was just delayed further.
In the former work, an additional refinement step is run using the better
orientation estimated from the longer segment obtained.
It is enough to completely detect most of the tested edges, but certainly
not all, especially if larger images with much longer edges are processed.
%The solution implemented in the former work was to let some arbitrary
%margin between the scan strip width and the assigned width to the detection,
%and to perform two fine detection steps, using for each of them the direction
%found at the former step.
This
process
could be itered as long as the blurred segment escapes from
This
operation
could be itered as long as the blurred segment escapes from
the directional scanner using as any fine detection steps as necessary.
But
the multiple detection of the same segment start
points pro
du
ces
a useless computational co
a
st.
But
at each iteration, already tested
points
are
proces
sed again,
thus producing
a useless computational cost.
Here the proposed solution is to dynamically a
dapt
the scan direction o
n
the
detection result
.
At each
posi
tion
$
i
+
1
$
, the scan strip is updated using the direction
of the blurred segment computed at
the former posi
tion
$
i
$
.
Here the proposed solution is to dynamically a
lign
the scan direction
t
o
the
blurred segment one all along the expansion stage
.
At each
itera
tion
$
i
+
1
$
, the scan strip is updated using the direction
of the blurred segment computed at
previous itera
tion
$
i
$
.
The adaptive directional scan
$
ADS
$
is then defined by :
\begin{equation}
%S_i = \mathcal{D}_{i-1} \cap \mathcal{N}_i
...
...
@@ -197,10 +201,14 @@ S_i = \mathcal{D}_i \cap \mathcal{N}_i \cap \mathcal{I}
\left
|
\begin{array}
{
l
}
\delta
(
\mathcal
{
N
}_
i) = -
\delta
^{
-1
}
(
\mathcal
{
D
}_
0)
\\
\cap
h
_
0(
\mathcal
{
N
}_
i) = h
_
0(
\mathcal
{
N
}_{
i-1
}
) + p(
\mathcal
{
D
}
)
\\
\cap
\mathcal
{
D
}_{
i+1
}
=
d
(
\mathcal
{
B
}_
i,
\varepsilon
)
\cap
\mathcal
{
D
}_{
i+1
}
=
D
(
\mathcal
{
B
}_
i,
\varepsilon
+ k
)
\end{array}
\right
.
\right\}
\end{equation}
The last clause expresses the update of the scan bounds at step
$
i
+
1
$
.
where
$
D
(
\mathcal
{
B
}_
i,w
)
$
is the scan strip aligned to the
detected segment at iteration
$
i
$
with width
$
w
$
.
In practice, the scan width is set a little greater than the assigned
width
$
\varepsilon
$
(
$
k
$
is a constant arbitrarily set to 4).
The last clause expresses the update of the scan bounds at iteration
$
i
+
1
$
.
Compared to static directional scans, the scan strip moves while
scan lines remain fixed.
An example of adaptive directional scan is given in
\RefFig
{
fig:adaption
}
.
...
...
@@ -225,8 +233,8 @@ An example of adaptive directional scan is given in \RefFig{fig:adaption}.
The assigned width
$
\varepsilon
$
to the blurred segment recognition algorithm
is initially set to a large value
$
\varepsilon
_
0
$
in order to allow the
detection of large blurred segments.
Then, when no more augmentation of the minimal width is observed a
s the
blurred segment expend
s (
$
\mu
_{
i
+
\lambda
}
=
\mu
_
i
$
), it is set to a much
Then, when no more augmentation of the minimal width is observed a
fter
$
\lambda
$
iteration
s (
$
\mu
_{
i
+
\lambda
}
=
\mu
_
i
$
), it is set to a much
stricter value able to circumscribe the possible interpretations of the
segment, that take into account the digitization margins:
\begin{equation}
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment