Image Processing Reference
In-Depth Information
which were constructed based on the orientation of the edges. Finally, a simple
remapping of each cell state to light intensity values is applied which is based on the
weighted sum of the neighboring light intensity values of every pixel. The method
manages to preserve the initial edges adequately while achieving high frame rates
for both color and grayscale images. In order to evaluate the performance of the pro-
posed method, a quantitative comparison with other related methods was applied
proving its effectiveness. More specifically, the presented experimental results in
terms of PSNR values and processing time prove the effectiveness of the proposed
method when compared with well-known methods as well as its suitability, espe-
cially for systems with low technical specifications when further image processing is
required.
The rest of the chapter is organized as follows. Section 2.2 provides the basic fun-
damentals of the CA as used in this chapter and the Canny edge detector principles.
In Sects. 2.3 and 2.4, an analysis of the proposed CA method and the experimental
results are presented, respectively. More specifically, experiments show that in both
visual comparisons and quantitative analysis, the results extracted by the proposed
CA based algorithm are better than those extracted from zero-ground, bilinear, bicu-
bic interpolation, near edge orientation. Finally, the hardware implementation of the
presented CA method for image resizing and conclusions are drawn in Sects. 2.5
and 2.6, respectively.
2.2
Cellular Automata Fundamentals and Canny Edge
Detector
2.2.1
CA Fundamentals
CA are decentralized, discrete space-time systems where interactions are local and
can be used to model physical systems [35]. A finite automaton could be defined by
the quadruple:
{
d
,
q
,
N
,
F
}
(2.1)
where
d
denotes the CA dimensions,
q
the total number of the used cell states,
N
the applied neighborhood and
f
the transition rules.
At each time step, every cell update its state based on its current state and the
state of its adjacent cells according to the defined transition set rule
f
.Itmustbe
highlighted that all cells of the CA grid update their state simultaneously leading to a
completely parallel system. The neighborhood of each cell is defined by variable
N
.
For a two-dimensional CA, two neighborhoods of range are often considered: Von
Neumann and Moore neighborhood. The Von Neumann neighborhood is a diamond
shaped neighborhood and can be used to define a set of cells surrounding a given
cell
(
x
0
,
y
0
)
. Equation 2.2 defines the Von Neumann neighborhood of range
r
.
N
N
(
x
0
,
y
0
)
=
{
(
x
,
y
)
:
|
x
−
x
0
|
+
|
y
−
y
0
|≤
(
r
)
}
(2.2)
Search WWH ::
Custom Search