Digital Signal Processing Reference
In-Depth Information
Show that the highpass filtering leads to the detection of edges in
the image.
(c) Calculate the 2D power spectrum of the filtered image. Comparing
this with the result of Problem 17.11 (d), highlight how the low-
frequency components have been attenuated in the filtered image.
17.15
Repeat Problem 17.14 for the 2D filter with the following impulse
response:
.
−
0
.
0442
0
0
0
−
0
.
3126
−
0
.
7815
−
0
.
3126
0
1
6
.
21
h
[
m
,
n
]
=
−
0
.
0442
−
0
.
7815
4
.
5532
−
0
.
7815
−
0
.
0442
0
−
0
.
3126
−
0
.
7815
−
0
.
3126
0
−
0
.
0442
0
0
17.16
Repeat Example 17.11 for the following selections of (4
4) pixels:
156
157
158
159
150
151
151
150
i
1
[
m
,
n
]
=
and
153
155
154
156
155
154
157
156
156
177
148
189
160
171
181
150
i
2
[
m
,
n
]
=
.
123
125
174
196
175
164
147
156
Show that the reconstruction error is greater for the second case, where
the neighboring pixels are less correlated.
17.17
Compress the image stored in the file “
lena.tif
” in the accompanying
CD using the JPEG standard with quality factors set to 80, 60, 40, 20,
and 10. Determine the compression ratio for different quality factors
and show that the subjective quality deteriorates as the quality factor is
decreased. Compute the mean square error for the compressed images.
Search WWH ::
Custom Search