Very Computer

by **tontok** » Fri, 08 Dec 2006 21:03:40

The pictures shown in the following URL;

http://www.unmannedspaceflight.com/index.php?act=Attach&type=post&id=...

are the original image of Vesta (left), its de-convoluted image (middle
/ focus depth:=3.8, iteration:=8 / software: Focus Corrector ) and
another de-convoluted image (right / focus depth:=4.3, iteration:=8 /
software: Focus Corrector).

For detail of Focus Corrector, visit;

http://139.134.5.123/tiddler2/c22508/focus.htm

by **ronvi..** » Sat, 09 Dec 2006 10:16:14

> The pictures shown in the following URL;

> http://www.unmannedspaceflight.com/index.php?act=Attach&type=post&id=...

> are the original image of Vesta (left), its de-convoluted image (middle
> / focus depth:=3.8, iteration:=8 / software: Focus Corrector ) and
> another de-convoluted image (right / focus depth:=4.3, iteration:=8 /
> software: Focus Corrector).

> For detail of Focus Corrector, visit;

> http://139.134.5.123/tiddler2/c22508/focus.htm

Wouldn't you need metadata (EXIF) to send shake (IS) data across to
make these adjustments? I do not think meta data contains that kind of
information. How do you get these beautiful results?

Thanks,
Ron

by **tontok** » Sat, 09 Dec 2006 22:01:17

I didn't use EXIF format file. I got the original image of Vesta from
HST website. It was a jpeg file and I converted it to bmp file for
dealing with it on my software (!)

My method is based on the inverse convolution (de-convolution) for
re-focusing blurred image. While for that purpose usually Fourier
transform is applied, my method carries out the inverse convolution
with iterative matrix calculations.

I have an online-gallery for some planetary images deconvolved by Focus
Corrector as,

http://www.geocities.com/q17320508/focuscorrector/gallery/planetary2/...

In the above gallery I fear some images might be over-deconvolved...

> > The pictures shown in the following URL;

> > http://www.unmannedspaceflight.com/index.php?act=Attach&type=post&id=...

> > are the original image of Vesta (left), its de-convoluted image (middle
> > / focus depth:=3.8, iteration:=8 / software: Focus Corrector ) and
> > another de-convoluted image (right / focus depth:=4.3, iteration:=8 /
> > software: Focus Corrector).

> > For detail of Focus Corrector, visit;

> > http://139.134.5.123/tiddler2/c22508/focus.htm

> Wouldn't you need metadata (EXIF) to send shake (IS) data across to
> make these adjustments? I do not think meta data contains that kind of
> information. How do you get these beautiful results?

> Thanks,
> Ron

by **Andrew Morto** » Sat, 09 Dec 2006 22:30:04

> I didn't use EXIF format file. I got the original image of Vesta from
> HST website. It was a jpeg file

Uh-oh!

Quote:> and I converted it to bmp file for
> dealing with it on my software (!)

> My method is based on the inverse convolution (de-convolution) for
> re-focusing blurred image. While for that purpose usually Fourier
> transform is applied, my method carries out the inverse convolution
> with iterative matrix calculations.

> I have an online-gallery for some planetary images deconvolved by
> Focus Corrector as,

> http://www.geocities.com/q17320508/focuscorrector/gallery/planetary2/...

> In the above gallery I fear some images might be over-deconvolved...

Definitely. Unfortunately, without knowing what the objects look like
without blur, we can't tell how good a job your algorithm is doing. How
about using more convenient objects, like trees or buildings, where you can
mount the camera on a tripod and take focused and de-focused images?

Andrew

by **Tom Nelso** » Sun, 10 Dec 2006 03:08:03

For those unwilling to spend $49.85 for Focus Corrector, you may find
the freeware Unshake a suitable alternative:
http://www.hamangia.freeserve.co.uk/Unshake/
It also uses Fourier transforms.
Tom Nelson
Tom Nelson Photography

> My method is based on the inverse convolution (de-convolution) for
> re-focusing blurred image. While for that purpose usually Fourier
> transform is applied, my method carries out the inverse convolution
> with iterative matrix calculations.
> > > For detail of Focus Corrector, visit;

> > > http://139.134.5.123/tiddler2/c22508/focus.htm

by **tontok** » Sun, 10 Dec 2006 21:29:13

Thank you for your suggestion. However to use the sample image taken by
myself might not be objective, since I may choose the example having
better result deliberately (!)

In the following webpage the comparison of deconvolved images of Tempel
1 processed by Photoshop and Focus Corrector is shown.

http://www.unmannedspaceflight.com/index.php?act=Attach&type=post&id=...

(Left: the original image taken by Deep Impact probe
Middle-left: processed by Photoshop
Middle-right: processed by Focus Corrector
Right: another genuine image taken at the position closer to the
surface of Tempel 1)

> > I didn't use EXIF format file. I got the original image of Vesta from
> > HST website. It was a jpeg file

> Uh-oh!

> > and I converted it to bmp file for
> > dealing with it on my software (!)

> > My method is based on the inverse convolution (de-convolution) for
> > re-focusing blurred image. While for that purpose usually Fourier
> > transform is applied, my method carries out the inverse convolution
> > with iterative matrix calculations.

> > I have an online-gallery for some planetary images deconvolved by
> > Focus Corrector as,

> > http://www.geocities.com/q17320508/focuscorrector/gallery/planetary2/...

> > In the above gallery I fear some images might be over-deconvolved...

> Definitely. Unfortunately, without knowing what the objects look like
> without blur, we can't tell how good a job your algorithm is doing. How
> about using more convenient objects, like trees or buildings, where you can
> mount the camera on a tripod and take focused and de-focused images?

> Andrew

by **ronvi..** » Mon, 11 Dec 2006 05:18:27

> For those unwilling to spend $49.85 for Focus Corrector, you may find
> the freeware Unshake a suitable alternative:
> http://www.hamangia.freeserve.co.uk/Unshake/
> It also uses Fourier transforms.
> Tom Nelson
> Tom Nelson Photography

Hi Tom,
The results you are getting are amazing but due to problems with my
computer I was unable to get it to work. I think it may be possible to
improve the user interface. I like the idea of an input folder and an
output folder. But if in addition to that if you generated a set of
output images, maybe twelve, according to a distribution (possibly
initially random) of input parameters then allow the user to select the
best, say two or three, of that set as a way to generate new sets of
parameters. That process could be repeated allowing the user do
eliminate by selection paths of failed parameters. It would also add
meaning to its name 'Small Wriggly Thing'.

Thank you for putting it together and good luck with your project,
Ron

Very Computer

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta

Unmasked Vesta