I'm looking for software that would help turn scanned images into bead
patterns.  Obviously any graphics program can resize an image down to any
size we want.  The problem with every graphics program I know, is that the
resizing begins at the upper left corner.  As a result, faces and other
images do not end up with their eyes or other details centered over an
appropriate set of pixels.

What I hope to find is a program that will let me define an appropriate
target grid and lay it over the original image to see a potential result.
In other words, I want it show me what happens if I shift the grid by a
fraction of a bead.  The ideal is to place the resizing grid anywhere
necessary to achieve a best result.

In addition to column/row grid patterns, the grid should support hexagonal
patterns.

In addition, the grid should support beads that are wider, or taller because
beads do not necessarily fit a square pattern.

In addition, the grid should be rotatable so that BOTH eyes, or other
details can be made to match up with an ideal set of pixel positions.

In addition to previewing a resizing, the preview should allow us to select
color reductions based on colors that we can preselect to match our
available beads or tiles.

Is there any such software?

Do you have a sample of what it is you are trying to achieve?

-- 
....Carl Frisk
Anger is a brief madness.
- Horace, 20 B.C.
http://www.carlfrisk.com


"WDS" <wdsnews.0505@oregoncity.com> wrote in message news:113vhmfl6lknh6d@corp.supernews.com...
> I'm looking for software that would help turn scanned images into bead
> patterns.  Obviously any graphics program can resize an image down to any
> size we want.  The problem with every graphics program I know, is that the
> resizing begins at the upper left corner.  As a result, faces and other
> images do not end up with their eyes or other details centered over an
> appropriate set of pixels.
> 
> What I hope to find is a program that will let me define an appropriate
> target grid and lay it over the original image to see a potential result.
> In other words, I want it show me what happens if I shift the grid by a
> fraction of a bead.  The ideal is to place the resizing grid anywhere
> necessary to achieve a best result.
> 
> In addition to column/row grid patterns, the grid should support hexagonal
> patterns.
> 
> In addition, the grid should support beads that are wider, or taller because
> beads do not necessarily fit a square pattern.
> 
> In addition, the grid should be rotatable so that BOTH eyes, or other
> details can be made to match up with an ideal set of pixel positions.
> 
> In addition to previewing a resizing, the preview should allow us to select
> color reductions based on colors that we can preselect to match our
> available beads or tiles.
> 
> Is there any such software?
> 
> 
>

WDS wrote:
> I'm looking for software that would help turn scanned images into
> bead patterns.  Obviously any graphics program can resize an
> image down to any size we want.  The problem with every graphics
> program I know, is that the resizing begins at the upper left
> corner.  As a result, faces and other images do not end up with
> their eyes or other details centered over an appropriate set of
> pixels.
>
> What I hope to find is a program that will let me define an
> appropriate target grid ... shift the grid by a fraction of a
> bead... the grid should support hexagonal patterns... should
> support beads that are wider, or taller because beads do not
> necessarily fit a square pattern... should be rotatable so that
> BOTH eyes, or other details can be made to match up with an ideal
> set of pixel positions.
>
> In addition to previewing a resizing, the preview should allow us
> to select color reductions based on colors that we can preselect
> to match our available beads or tiles.

What a fascinating application! Although not automatic, PSP has the 
tools to do all of this except combining the preview and the color 
reduction. Perhaps someone will come by with a way to do that, too.

Experiment with a two-layer image. Open a working copy of a scanned 
image. Promote the background layer and increase the canvas size to 
make room for sliding the image around. Add a new layer containing 
the grid. For the experiment, the supplied Gridmaker script will 
make the new layer and a rectangular grid. Hexagonal grids can be 
done too, but are not supplied.

To make the resizing hit the right pixels, move and rotate the image 
rather than the grid. Select the image layer in the Layers palette 
and use the Deform tool to move and rotate the image.

For the preview, Window > New Window. Use the Zoom tool to shrink 
the new window to the ultimate size. Changes in the first window 
will show in the new small window. Turn the visibility of the grid 
layer on and off as needed to judge the result.

When it is ready, turn off the grid and resize the image--probably 
to the same percentage you were watching in the zoomed-out preview 
window.

PSP does color reduction very well. If you have a color palette 
already, simply Image > Palette > Load Palette. To make a palette 
from an image that you have on hand, Image > Decrease Color Depth. 
Choose the number of colors and reduction method, do it, and save 
the palette with Image > Palette > Save Palette.

With its unlimited undo, PSP lets you back up to tweak an earlier 
step easily.
-- 
Fred Hiltz,  fhiltz at yahoo dot com
C-Tech volunteer 

Carl Frisk wrote:

> Do you have a sample of what it is you are trying to achieve?

You know, Carl, if Canopus, Tim, Antonio, Jules, Kiri and Kris were as 
nice as you, this group would be great!!!!

:-)

Joe Cilinceon is nice, too. But Kris doesn't like him anymore, since he 
stopped upgrading PSP!

:-)

Uni

Fred,

Thank you.  Your reply goes beyond the call of duty.

I agree this would make a fascinating new application.  I've written enough
software to know this wouldn't be real hard, assuming you have something
like PSP to begin with.  I confess most users might not know enough to begin
with a small or medium size scanned image, and suffer from slow performance
as we attempt to simultaneously rotate, resize, color reduce and preview in
real-time.  But the new program should be smart enough to automatically
create a smaller version of the original in order to create a
fast-yet-accurate preview.  The previewer should be smart enough to follow
the "instant preview" with a more accurate preview if the user paused long
enough to see it.

Thank you for the procedure you gave me.  I will experiment with it.




"Fred Hiltz" <not@home.ca> wrote in message
news:IuSdna-gzcj_hN3fRVn-pA@adelphia.com...
> WDS wrote:
> > I'm looking for software that would help turn scanned images into
> > bead patterns.  Obviously any graphics program can resize an
> > image down to any size we want.  The problem with every graphics
> > program I know, is that the resizing begins at the upper left
> > corner.  As a result, faces and other images do not end up with
> > their eyes or other details centered over an appropriate set of
> > pixels.
> >
> > What I hope to find is a program that will let me define an
> > appropriate target grid ... shift the grid by a fraction of a
> > bead... the grid should support hexagonal patterns... should
> > support beads that are wider, or taller because beads do not
> > necessarily fit a square pattern... should be rotatable so that
> > BOTH eyes, or other details can be made to match up with an ideal
> > set of pixel positions.
> >
> > In addition to previewing a resizing, the preview should allow us
> > to select color reductions based on colors that we can preselect
> > to match our available beads or tiles.
>
> What a fascinating application! Although not automatic, PSP has the
> tools to do all of this except combining the preview and the color
> reduction. Perhaps someone will come by with a way to do that, too.
>
> Experiment with a two-layer image. Open a working copy of a scanned
> image. Promote the background layer and increase the canvas size to
> make room for sliding the image around. Add a new layer containing
> the grid. For the experiment, the supplied Gridmaker script will
> make the new layer and a rectangular grid. Hexagonal grids can be
> done too, but are not supplied.
>
> To make the resizing hit the right pixels, move and rotate the image
> rather than the grid. Select the image layer in the Layers palette
> and use the Deform tool to move and rotate the image.
>
> For the preview, Window > New Window. Use the Zoom tool to shrink
> the new window to the ultimate size. Changes in the first window
> will show in the new small window. Turn the visibility of the grid
> layer on and off as needed to judge the result.
>
> When it is ready, turn off the grid and resize the image--probably
> to the same percentage you were watching in the zoomed-out preview
> window.
>
> PSP does color reduction very well. If you have a color palette
> already, simply Image > Palette > Load Palette. To make a palette
> from an image that you have on hand, Image > Decrease Color Depth.
> Choose the number of colors and reduction method, do it, and save
> the palette with Image > Palette > Save Palette.
>
> With its unlimited undo, PSP lets you back up to tweak an earlier
> step easily.
> -- 
> Fred Hiltz,  fhiltz at yahoo dot com
> C-Tech volunteer
>

Fred I still don't see the goal of the application.  Perhaps it's a linguistics problem.  Could you explain the end 
result or show a sample as this does sound interesting?

-- 
....Carl Frisk
Anger is a brief madness.
- Horace, 20 B.C.
http://www.carlfrisk.com


"Fred Hiltz" <not@home.ca> wrote in message news:IuSdna-gzcj_hN3fRVn-pA@adelphia.com...
> WDS wrote:
>> I'm looking for software that would help turn scanned images into
>> bead patterns.
<*SNIP*>
>
> What a fascinating application! Although not automatic, PSP has the
> tools to do all of this except combining the preview and the color
> reduction. Perhaps someone will come by with a way to do that, too.
>
> Experiment with a two-layer image. Open a working copy of a scanned
> image. Promote the background layer and increase the canvas size to
> make room for sliding the image around. Add a new layer containing
> the grid. For the experiment, the supplied Gridmaker script will
> make the new layer and a rectangular grid. Hexagonal grids can be
> done too, but are not supplied.
>
> To make the resizing hit the right pixels, move and rotate the image
> rather than the grid. Select the image layer in the Layers palette
> and use the Deform tool to move and rotate the image.
>
> For the preview, Window > New Window. Use the Zoom tool to shrink
> the new window to the ultimate size. Changes in the first window
> will show in the new small window. Turn the visibility of the grid
> layer on and off as needed to judge the result.
>
> When it is ready, turn off the grid and resize the image--probably
> to the same percentage you were watching in the zoomed-out preview
> window.
>
> PSP does color reduction very well. If you have a color palette
> already, simply Image > Palette > Load Palette. To make a palette
> from an image that you have on hand, Image > Decrease Color Depth.
> Choose the number of colors and reduction method, do it, and save
> the palette with Image > Palette > Save Palette.
>
> With its unlimited undo, PSP lets you back up to tweak an earlier
> step easily.
> -- 
> Fred Hiltz,  fhiltz at yahoo dot com
> C-Tech volunteer
> 

WDS wrote:
> 
> I'm looking for software that would help turn scanned images into bead
> patterns.  Obviously any graphics program can resize an image down to any
> size we want.  The problem with every graphics program I know, is that the
> resizing begins at the upper left corner.

That's not a problem.

> As a result, faces and other
> images do not end up with their eyes or other details centered over an
> appropriate set of pixels.

Put your mouse of the center of each eye in turn. Write down the
X and Y coordinates. Decide how many beads will fit center to
center between these coordinates. The bead size becomes your
unit of measure. If the leftmost and topmost eye coordinate is
(X1, Y1), compute how many bead sizes you can subtract from each
without going less than zero. Let's assume that after subtracting
the necessary number of beads the coordinates are reduced to
(X2, Y2). All you need to do now is crop off a piece of image that
is the full height and from 0 to X2 in width. Then crop off a
piece of image that is the full width and from 0 to Y2 in height.
In other words, you remove a fraction of a bead width from the
left and top of the image so that there is a whole number of
beads from the edge of the image to the eye position in the new
image. Now resizing will work just fine.

> What I hope to find is a program that will let me define an appropriate
> target grid and lay it over the original image to see a potential result.

Why do it by guesswork when you can just do the math?

> In other words, I want it show me what happens if I shift the grid by a
> fraction of a bead.  The ideal is to place the resizing grid anywhere
> necessary to achieve a best result.

If you want want to play around with grids, run the GridMaker
script and move the resulting grid layer around until you are
happy. Then crop to the edge of the grid or determine the
relevant pixel offsets.

> In addition to column/row grid patterns, the grid should support hexagonal
> patterns.

The GridMaker script supports only a rectangular grid. However,
the same principle - flood filling a new layer with a grid cell
having a transparent background - can be used to create a
hexagonal grid (or a range of other grids, provided the cells
are shaped to tile).

> In addition, the grid should support beads that are wider, or taller because
> beads do not necessarily fit a square pattern.

Wider or taller than what? Once you have a grid layer overlay
you can resize it how you want, using different scalings in the
horizontal and vertical directions. You can do this numerically
with Image > Resize or by eyeballing with the Deform tool. The
Deform tool will let you rotate the grid and apply a skew or
perspective deformation as well.

> In addition, the grid should be rotatable so that BOTH eyes, or other
> details can be made to match up with an ideal set of pixel positions.

There you go, my previous paragraph anticipates your needs :)
 
> In addition to previewing a resizing, the preview should allow us to select
> color reductions based on colors that we can preselect to match our
> available beads or tiles.

Do a color reduction then edit the palette to replace the colors
in it with the most similar respective bead color. You can also
load a standard bead color palette into the paletted image and
use nearest color matching to replace the image colors with bead
colors.

> Is there any such software?

There's PSP, the subject of the newsgroup to which you are
posting. Various of the processes I described can be scripted in
PSP so, since you are a programmer, that should be easy. On top
of that, PSP will actually render beads (see the Balls & Bubbles
filter) for which you can also use scripting.

Carl Frisk wrote:
> Fred I still don't see the goal of the application.  Perhaps it's
> a linguistics problem.  Could you explain the end
> result or show a sample as this does sound interesting?

Hi Carl. I could explain my imperfect understanding, but perhaps WDS 
will rejoin the thread with an example or two. I envisioned a 
photograph applied to a flat layer of bead-work in such a way that 
important features of the photo hit the centers of beads and are not 
lost in the cracks between.
-- 
Fred Hiltz,  fhiltz at yahoo dot com
C-Tech volunteer 

> Hi Carl. I could explain my imperfect understanding, but
> perhaps WDS will rejoin the thread with an example or two.

First, I truly appreciate the suggestions that have been given.  These
techniques are far superior to what I've been doing.

To answer your question, the software developer's challenge is to show the
user an end-result in real-time.  We need to move, rotate, size,
color-match, and color-reduce in continuous real-time operations.

I work with wood and ceramic tiles.  My daughters work with beads.  All
three hobbies can benefit from using the computer to generate accurate
patterns.  For example, the darkness of your available wood can make an
extreme difference in the accuracy of your color reduction.  (Is the medium
shade area larger or smaller?)

The end-user's challenge is trying to fit the available space while trying
to figure out which details matter.  Often times a smaller result will
provide a better result simply because of the way the details fit on the
pattern.

I mentioned eyes as my example because they often matter.  But other
details, such as edge shape, could be just as significant for creating a
recognizable result.  Only a real-time preview can let you discover all the
details that should be considered.  The reason for the real-time rotate
requirement, is that it promises the best potential for discovering which
details can be revealed.  (How many beads should we include in the bottom of
the curve?  Should we try to accomodate other details?  how many?)

The suggestions you gave me in this forum are very helpful.  Thank you.  But
until we're able to move, rotate, size, color-match, and color-reduce in
continuous real-time operations, we won't have fulfilled the potential the
computer promises for our hobbies.   In the mean-time, experts like you in
this forum help it to keep getting better.

Thanks again.


"Fred Hiltz" <not@home.ca> wrote in message
news:H6udneXIbKIJ7NrfRVn-hQ@adelphia.com...
> Carl Frisk wrote:
> > Fred I still don't see the goal of the application.  Perhaps it's
> > a linguistics problem.  Could you explain the end
> > result or show a sample as this does sound interesting?
>
> Hi Carl. I could explain my imperfect understanding, but perhaps WDS
> will rejoin the thread with an example or two. I envisioned a
> photograph applied to a flat layer of bead-work in such a way that
> important features of the photo hit the centers of beads and are not
> lost in the cracks between.
> -- 
> Fred Hiltz,  fhiltz at yahoo dot com
> C-Tech volunteer
>

Thanks Fred,
  I was thinking in a different direction.  I've actually written something like this about 5 years ago for a similar 
application.

-- 
....Carl Frisk
Anger is a brief madness.
- Horace, 20 B.C.
http://www.carlfrisk.com


"Fred Hiltz" <not@home.ca> wrote in message news:H6udneXIbKIJ7NrfRVn-hQ@adelphia.com...
> Carl Frisk wrote:
>> Fred I still don't see the goal of the application.  Perhaps it's
>> a linguistics problem.  Could you explain the end
>> result or show a sample as this does sound interesting?
>
> Hi Carl. I could explain my imperfect understanding, but perhaps WDS
> will rejoin the thread with an example or two. I envisioned a
> photograph applied to a flat layer of bead-work in such a way that
> important features of the photo hit the centers of beads and are not
> lost in the cracks between.
> -- 
> Fred Hiltz,  fhiltz at yahoo dot com
> C-Tech volunteer
> 

Now I see.  I was thinking in a different direction.  Which is why I asked.  Thank you for your answer.

It seems that what you are looking for is a program that will take a photo/image and simulate what it will look like in 
real life as laid out across a medium.  A medium such as beads, ceramic tiles, wood.  So that you can adjust it in real 
time to make it look right and satisfy the customers taste.  This would then provide a template you could work from to 
produce the final product/craft.

I was approached about 5 years ago to write just such a program.  Though the application was for an industrial use not 
related to yours.

PSP could be adapted with scripting to accomplish this goal.



-- 
....Carl Frisk
Anger is a brief madness.
- Horace, 20 B.C.
http://www.carlfrisk.com


"WDS" <wdsnews.0505@oregoncity.com> wrote in message news:114grta16j5gm07@corp.supernews.com...
>> Hi Carl. I could explain my imperfect understanding, but
>> perhaps WDS will rejoin the thread with an example or two.
>
> First, I truly appreciate the suggestions that have been given.  These
> techniques are far superior to what I've been doing.
>
> To answer your question, the software developer's challenge is to show the
> user an end-result in real-time.  We need to move, rotate, size,
> color-match, and color-reduce in continuous real-time operations.
>
> I work with wood and ceramic tiles.  My daughters work with beads.  All
> three hobbies can benefit from using the computer to generate accurate
> patterns.  For example, the darkness of your available wood can make an
> extreme difference in the accuracy of your color reduction.  (Is the medium
> shade area larger or smaller?)
>
> The end-user's challenge is trying to fit the available space while trying
> to figure out which details matter.  Often times a smaller result will
> provide a better result simply because of the way the details fit on the
> pattern.
>
> I mentioned eyes as my example because they often matter.  But other
> details, such as edge shape, could be just as significant for creating a
> recognizable result.  Only a real-time preview can let you discover all the
> details that should be considered.  The reason for the real-time rotate
> requirement, is that it promises the best potential for discovering which
> details can be revealed.  (How many beads should we include in the bottom of
> the curve?  Should we try to accomodate other details?  how many?)
>
> The suggestions you gave me in this forum are very helpful.  Thank you.  But
> until we're able to move, rotate, size, color-match, and color-reduce in
> continuous real-time operations, we won't have fulfilled the potential the
> computer promises for our hobbies.   In the mean-time, experts like you in
> this forum help it to keep getting better.
>
> Thanks again.
>
>
> "Fred Hiltz" <not@home.ca> wrote in message
> news:H6udneXIbKIJ7NrfRVn-hQ@adelphia.com...
>> Carl Frisk wrote:
>> > Fred I still don't see the goal of the application.  Perhaps it's
>> > a linguistics problem.  Could you explain the end
>> > result or show a sample as this does sound interesting?
>>
>> Hi Carl. I could explain my imperfect understanding, but perhaps WDS
>> will rejoin the thread with an example or two. I envisioned a
>> photograph applied to a flat layer of bead-work in such a way that
>> important features of the photo hit the centers of beads and are not
>> lost in the cracks between.
>> -- 
>> Fred Hiltz,  fhiltz at yahoo dot com
>> C-Tech volunteer
>>
>
>