Digital Signal Processing Reference
In-Depth Information
For real world, the received video quality is measured by reference objects. For
vector in feature space, vector-valued image and its popularity is mainly due to its
present in space. The definition of special exponent distribution is observed in the
vector space. For Gabor space, the different channels give special distribution in high
dimension space. We can define distance between vectors as
MM
d
=
||
x
−
y
||
2
(1)
ij
,
i
j
i
==
11
j
where
x
is a value in vector
X
. Based on this measure, distribution between vec-
tors in Gabor space is computed and filtered.
An input image is filtered with filters with various frequency and orientation at a
channel. We can analysis these paralleling channels. For the customized regions have
some common properties at different channels, we can separate special regions using
there relationship of channels. The proposed paradigm is shown in Fig.2.
Fig. 2.
Proposed Gabor Filters Process paradigm
The analysis can be performed in each filter channel using different scale and the-
ta. It is possible to label customized regions in an out-put image.
f=5.2632, th eta=0
f= 3.9370 , theta=45
f=4.2017, theta=90
f=3.2258, theta=135
Fig. 3.
The filtering results of Gabor
