RSS News FeedIs there any way to enhance this image
Message-ID:<f7d49b62-d5db-4feb-9e98-5ff3d77a21d0@41g2000yqf.googlegroups.com>
Subject:
Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 13:36:04 +0100
Hello, It looks like there are some sort of ripples around image's edges... <a target="_blank" href="http://g.imagehost.org/download/0831/ whitehouse">whitehouse.tiff (73 KB)</a> I'm out of ideas how to enhance it, so any idea how to get rid of those ripples would be appreciated, Thanks
Message-ID:<faf5ff04-b115-4116-968c-d31e155bbace@f40g2000pri.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 14:32:19 +0100
On 2$B7n(B7$BF|(B, $B2<8a(B8$B;~(B36$BJ,(B, stasgold <stasg...@gmail.com> wrote: > Hello, > > It looks like there are some sort of ripples around image's edges... > <a target="_blank" href="http://g.imagehost.org/download/0831/ > whitehouse">whitehouse.tiff (73 KB)</a> > > I'm out of ideas how to enhance it, so any idea how to get rid of > those ripples would be appreciated, > > Thanks interesting... Seems to me that this picture had been manipulated (attenuation?) on the Fourier domain within a ca. 44 by 44 Fourier-pixel-square centered at DC according to an inspection in the Fourier domain. But I do not have an idea how this could be recovered ... I've tried taking out the Fourier vectors <0,10> & <0,11> to see some change what could not be called an enhancement....
Message-ID:<c7a42bb3-df86-401e-b1f6-b1e6f602df6a@13g2000yql.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sun, 8 Feb 2009 18:47:54 +0100
> interesting... =A0Seems to me that this picture had been manipulated > (attenuation?) on the Fourier domain within a ca. 44 by 44 > Fourier-pixel-square centered at DC according to an inspection > in the Fourier domain. =A0But I do not have an idea how this could be > recovered ... I've tried taking out the Fourier vectors <0,10> & > <0,11> > to see some change what could not be called an enhancement.... Thanks, this square around DC indeed looks artificial ( how I haven't noted its existence ? :) I tried to lower its values by different factors, factor of 2 , it seems, does the job
Message-ID:<498ddb87$1@news.arcor-ip.de>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 20:05:40 +0100
what about some nonlinear bandstop.
try this in matlab or octave:
>Id = im2double(imread('whitehouse.jpg'));
>lowpass = 21;
>highpass = 7;
>Idm = medfilt2(Id, [lowpass lowpass]);
>Ide = Id - medfilt2(Id, [highpass highpass]);
>IdX = Idm + Ide;
>imshow(IdX);
greets
Message-ID:<a028d376-529d-4d9f-9705-cbb184ea9367@x9g2000yqk.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 23:33:15 +0100
Nice work Johann! That does an awesome job, in my opinion.
Message-ID:<d613114c-8951-42ad-8748-327c4f791a34@f20g2000yqg.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sun, 8 Feb 2009 18:52:04 +0100
On Feb 7, 9:05=A0pm, Johann Blaser <johann.bla...@gmx.de> wrote:
> what about some nonlinear bandstop.
>
> try this in matlab or octave:
>
> =A0>Id =3D im2double(imread('whitehouse.jpg'));
> =A0>lowpass =3D 21;
> =A0>highpass =3D 7;
> =A0>Idm =3D medfilt2(Id, [lowpass lowpass]);
> =A0>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> =A0>IdX =3D Idm + Ide;
> =A0>imshow(IdX);
>
> greets
Thanks for your reply, Johann
I thought median filter is useful in removing salt & pepper noise
only,
but your code makes wonders removing those ripples
Message-ID:<d9424f5d-535b-4a31-8f81-a5c36a63e5a1@e3g2000vbe.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Fri, 20 Feb 2009 22:54:17 +0100
On Feb 7, 2:05=A0pm, Johann Blaser <johann.bla...@gmx.de> wrote:
> what about some nonlinear bandstop.
>
> try this in matlab or octave:
>
> =A0>Id =3D im2double(imread('whitehouse.jpg'));
> =A0>lowpass =3D 21;
> =A0>highpass =3D 7;
> =A0>Idm =3D medfilt2(Id, [lowpass lowpass]);
> =A0>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> =A0>IdX =3D Idm + Ide;
> =A0>imshow(IdX);
>
> greets
Johann, it is nice to see your code. Can you explain a little bit more
on the code above? why you let
>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> IdX =3D Idm + Ide;
to make it work?
I am new to image processing, and I thought median filter is to remove
noise, like I always use img_fil =3D medfilt2(img, [windsize, windsize])
and don't know I can do something on subtracting the filtered image. I
would like to learn more from your explaination so that I can know
when I can do like that.
Thanks in advance for your further explaination.
Message-ID:<49a91cf4$1@news.arcor-ip.de>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 28 Feb 2009 12:16:00 +0100
Hi Xiaoxiao, it's just a way to get rid of image-parts that have certain frequencies. When i subtract a light lowpass filtered signal from the original signal, i get a strong highpass filtered signal. When i then add this signal to a strong lowpass filtered signal, i get a bandstop filter, because from the difference of two "lowpasses", something drops out. To a certain extent, median filtering does the the same as a linear filter. If you rank up a sinusoid signal, like the ripple in the image and take the median, you get something near the mean, as long as the frequency is high enough. If the frequency is too low, the mean departs too much from the median, and the filter will produce artefacts. Kind regards, Johann
Message-ID:<f7d49b62-d5db-4feb-9e98-5ff3d77a21d0@41g2000yqf.googlegroups.com>
Subject:
Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 13:36:04 +0100
Hello, It looks like there are some sort of ripples around image's edges... <a target="_blank" href="http://g.imagehost.org/download/0831/ whitehouse">whitehouse.tiff (73 KB)</a> I'm out of ideas how to enhance it, so any idea how to get rid of those ripples would be appreciated, Thanks
Message-ID:<faf5ff04-b115-4116-968c-d31e155bbace@f40g2000pri.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 14:32:19 +0100
On 2$B7n(B7$BF|(B, $B2<8a(B8$B;~(B36$BJ,(B, stasgold <stasg...@gmail.com> wrote: > Hello, > > It looks like there are some sort of ripples around image's edges... > <a target="_blank" href="http://g.imagehost.org/download/0831/ > whitehouse">whitehouse.tiff (73 KB)</a> > > I'm out of ideas how to enhance it, so any idea how to get rid of > those ripples would be appreciated, > > Thanks interesting... Seems to me that this picture had been manipulated (attenuation?) on the Fourier domain within a ca. 44 by 44 Fourier-pixel-square centered at DC according to an inspection in the Fourier domain. But I do not have an idea how this could be recovered ... I've tried taking out the Fourier vectors <0,10> & <0,11> to see some change what could not be called an enhancement....
Message-ID:<c7a42bb3-df86-401e-b1f6-b1e6f602df6a@13g2000yql.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sun, 8 Feb 2009 18:47:54 +0100
> interesting... =A0Seems to me that this picture had been manipulated > (attenuation?) on the Fourier domain within a ca. 44 by 44 > Fourier-pixel-square centered at DC according to an inspection > in the Fourier domain. =A0But I do not have an idea how this could be > recovered ... I've tried taking out the Fourier vectors <0,10> & > <0,11> > to see some change what could not be called an enhancement.... Thanks, this square around DC indeed looks artificial ( how I haven't noted its existence ? :) I tried to lower its values by different factors, factor of 2 , it seems, does the job
Message-ID:<498ddb87$1@news.arcor-ip.de>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 20:05:40 +0100
what about some nonlinear bandstop.
try this in matlab or octave:
>Id = im2double(imread('whitehouse.jpg'));
>lowpass = 21;
>highpass = 7;
>Idm = medfilt2(Id, [lowpass lowpass]);
>Ide = Id - medfilt2(Id, [highpass highpass]);
>IdX = Idm + Ide;
>imshow(IdX);
greets
Message-ID:<a028d376-529d-4d9f-9705-cbb184ea9367@x9g2000yqk.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 7 Feb 2009 23:33:15 +0100
Nice work Johann! That does an awesome job, in my opinion.
Message-ID:<d613114c-8951-42ad-8748-327c4f791a34@f20g2000yqg.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Sun, 8 Feb 2009 18:52:04 +0100
On Feb 7, 9:05=A0pm, Johann Blaser <johann.bla...@gmx.de> wrote:
> what about some nonlinear bandstop.
>
> try this in matlab or octave:
>
> =A0>Id =3D im2double(imread('whitehouse.jpg'));
> =A0>lowpass =3D 21;
> =A0>highpass =3D 7;
> =A0>Idm =3D medfilt2(Id, [lowpass lowpass]);
> =A0>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> =A0>IdX =3D Idm + Ide;
> =A0>imshow(IdX);
>
> greets
Thanks for your reply, Johann
I thought median filter is useful in removing salt & pepper noise
only,
but your code makes wonders removing those ripples
Message-ID:<d9424f5d-535b-4a31-8f81-a5c36a63e5a1@e3g2000vbe.googlegroups.com>
Subject:
Re: Is there any way to enhance this image?
Date:Fri, 20 Feb 2009 22:54:17 +0100
On Feb 7, 2:05=A0pm, Johann Blaser <johann.bla...@gmx.de> wrote:
> what about some nonlinear bandstop.
>
> try this in matlab or octave:
>
> =A0>Id =3D im2double(imread('whitehouse.jpg'));
> =A0>lowpass =3D 21;
> =A0>highpass =3D 7;
> =A0>Idm =3D medfilt2(Id, [lowpass lowpass]);
> =A0>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> =A0>IdX =3D Idm + Ide;
> =A0>imshow(IdX);
>
> greets
Johann, it is nice to see your code. Can you explain a little bit more
on the code above? why you let
>Ide =3D Id - medfilt2(Id, [highpass highpass]);
> IdX =3D Idm + Ide;
to make it work?
I am new to image processing, and I thought median filter is to remove
noise, like I always use img_fil =3D medfilt2(img, [windsize, windsize])
and don't know I can do something on subtracting the filtered image. I
would like to learn more from your explaination so that I can know
when I can do like that.
Thanks in advance for your further explaination.
Message-ID:<49a91cf4$1@news.arcor-ip.de>
Subject:
Re: Is there any way to enhance this image?
Date:Sat, 28 Feb 2009 12:16:00 +0100
Hi Xiaoxiao, it's just a way to get rid of image-parts that have certain frequencies. When i subtract a light lowpass filtered signal from the original signal, i get a strong highpass filtered signal. When i then add this signal to a strong lowpass filtered signal, i get a bandstop filter, because from the difference of two "lowpasses", something drops out. To a certain extent, median filtering does the the same as a linear filter. If you rank up a sinusoid signal, like the ripple in the image and take the median, you get something near the mean, as long as the frequency is high enough. If the frequency is too low, the mean departs too much from the median, and the filter will produce artefacts. Kind regards, Johann