Image analysis – Image enhancement or restoration – Image filter
Reexamination Certificate
2000-07-28
2004-09-21
Ahmed, Samir (Department: 2623)
Image analysis
Image enhancement or restoration
Image filter
C382S266000, C348S470000, C348S533000, C348S610000, C348S627000, C375S240290
Reexamination Certificate
active
06795588
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a noise detector and a noise detection method for detecting noise of an image signal as well as a signal processor and a signal processing method employing the noise detector and the noise detection method.
2. Description of the Prior Art
For example, Japanese Patent Laying-Open No. 7-307942 discloses a conventional signal processor eliminating noise from an image signal.
FIG. 5
is a block diagram showing the structure of the conventional signal processor.
The signal processor shown in
FIG. 5
comprises a vertical adaptive filter
8
and a horizontal adaptive filter
9
. The vertical adaptive filter
8
includes a vertical logical filter
41
, a vertical unevenness determination part
42
and a selector
43
. The horizontal adaptive filter
9
includes a horizontal logical filter
44
, a horizontal unevenness determination part
45
and a selector
46
.
When an image signal is input, the vertical unevenness determination part
42
determines the magnitude of vertical unevenness around each pixel and outputs the result of the determination to the selector
43
. When the vertical unevenness determination part
42
determines that the vertical unevenness around the pixel to be processed is large, the selector
43
outputs a signal of the pixel to the horizontal adaptive filter
9
as such. When the vertical unevenness determination part
42
determines that the vertical unevenness is small, the selector
43
outputs the image signal from which vertical noise is eliminated by the vertical logical filter
41
to the horizontal adaptive filter
9
.
Then, the horizontal unevenness determination part
45
determines the magnitude of horizontal unevenness around each pixel and outputs the result of the determination to the selector
46
. When the horizontal unevenness determination part
45
determines that the horizontal unevenness around the pixel to be processed is large, the selector
46
outputs the image signal output from the vertical adaptive filter
8
as such. When the horizontal unevenness determination part
45
determines that the horizontal unevenness is small, the selector
46
outputs the image signal from which horizontal noise is eliminated by the horizontal logical filter
44
.
Through the aforementioned operations, the signal processor shown in
FIG. 5
can eliminate noise from segments or edges included in the image signal without deteriorating the resolution, i.e., without blurring the segments or edges.
While the aforementioned conventional signal processor can reproduce large-amplitude edges without blurring the same, however, the vertical and horizontal logical filters
41
and
44
disadvantageously blur portions continuously exhibiting fine details, such as the surfaces of clothes, the skin of a human being and a wood, for example, specific to the image signal and reduce the texture thereof.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a signal processor and a signal processing method capable of properly correcting an image signal without reducing the texture specific to the image signal also in a portion continuously exhibiting fine details.
Another object of the present invention is to provide a noise detector and a noise detection method, employed for the aforementioned signal processor and signal processing method, capable of detecting mosquito noise and ringing of an image signal.
A noise detector according to an aspect of the present invention comprises a high-frequency component extraction circuit extracting a high-frequency component of an input image signal, a level detection circuit detecting the level of the high-frequency component extracted by the high-frequency component extraction circuit and outputting the result of detection as a level signal and a noise detection circuit detecting a small-amplitude edge around a large-amplitude edge of the image signal as noise on the basis of the level signal detected by the level detection circuit.
The noise detector extracts the high-frequency component of the image signal and detects the level thereof, whereby the small-amplitude edge around the large-amplitude edge of the image signal can be detected on the basis of the detected level and this portion can be detected as noise. Consequently, mosquito noise and ringing caused around the large-amplitude edge of the image signal can be detected.
The noise detection circuit may include an extension circuit outputting an extended signal obtained by extending the peak of the level signal detected by the level detection circuit, a conversion circuit outputting a 1
signal obtained by converting the level of the extended signal output from the extension circuit to 1
times and a small-amplitude detection circuit comparing the 1
signal output from the conversion circuit with the level signal output from the level detection circuit for detecting the small-amplitude edge around the large-amplitude edge of the image signal.
In this case, the small-amplitude edge, existing around the large-amplitude edge corresponding to the peak of the level signal, smaller than 1
of the level of the large-amplitude edge can be detected by comparing the 1
signal obtained by extending the peak of the level signal and compressing the same to 1
times with the level signal, and the small-amplitude edge can be detected as mosquito noise and ringing.
The noise detection circuit may further include a large-amplitude detection circuit comparing the extended signal output from the extension circuit with a prescribed reference value for detecting a large-amplitude edge having a greater level than the reference value, for detecting the small-amplitude edge detected by the small-amplitude detection circuit as noise with respect to the large-amplitude edge detected by the large-amplitude detection circuit.
In this case, the large-amplitude edge, which is a signal specific to the image signal having a greater level than the reference value can be detected by comparing the extended signal obtained by extending the peak of the level signal corresponding to the large-amplitude edge with the prescribed reference value. Therefore, a small-amplitude edge around the large-amplitude edge, which is a signal specific to the image signal, can be detected as noise, and mosquito noise and ringing can be more correctly detected.
The small-amplitude detection circuit may output a first detection result signal indicating that the 1
signal output from the conversion circuit is greater than the level signal output from the level detection circuit, the large-amplitude detection circuit may output a second detection result signal indicating that the extended signal output from the extension circuit is greater than the reference value, and the noise detection circuit may further include a logic circuit ANDing the first detection result signal output from the small-amplitude detection circuit and the second detection result signal output from the large-amplitude detection circuit.
In this case, the first detection result signal indicates a portion where the 1
signal is greater than the level signal, i.e., a portion excluding a large-amplitude edge portion while the second detection result signal indicates a portion where the extended signal is greater than the reference value, i.e., a small-amplitude edge portion around the large-amplitude edge including the large-amplitude edge portion. Therefore, only the small-amplitude edge around the large-amplitude edge can be detected by ANDing the outputs. Consequently, only mosquito noise and ringing can be detected from the portion around the large-amplitude edge of the image signal with a simple structure employing a binary signal.
The high-frequency component extraction circuit may include a horizontal extraction circuit extracting a horizontal high-frequency component of the input image signal, a vertical extraction circuit extracting a vertical high-frequency component of the input image signal and an adder circuit adding the horizontal hi
Nio Yutaka
Okumura Naoji
Tanaka Kazuto
Terai Katsumi
Ahmed Samir
Bhatnagar Anand
Matsushita Electrical Industrial Co. Ltd.
Westerman Hattori Daniels & Adrian LLP
LandOfFree
Noise detector, noise detection method, signal processor and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Noise detector, noise detection method, signal processor and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Noise detector, noise detection method, signal processor and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3264751