Board index » delphi » 3D picture

2003-07-25 02:43:23 PM
delphi281
Does anyone have any information what 3D picture is?
If i have a bmp of some object from left view and from right view,
can i combine those somehow to get something 3D? And will it
improve viewing experience of this object?
Thanx.
 

George writes:
Try this link ...
www.stereoscopy.com/3d-info/
--
Hope this helps,
Regards,
Ananth B.
 

Thanx Ananth, but I'd like to see some Delphi code...
"Ananth B." <ananth__b-at-hotmail.com>writes

George
Try this: Take the left part of the picture and mask out green and blue,
take the right part and mask out red color and combine both together.
NewPixel[x,y] := LeftPixel[x,y].rgbRed + RightPixel[x,y].rgbGreen +
RightPixel[x,y].rgbBlue
You will get a anaglyphic image to be viewed with a red-cyan colored 3d
glass.
Good luck!
"George" <XXXX@XXXXX.COM>schrieb im Newsbeitrag

IMHO you cant view 3D pictures without glasses or a special (expensive) 3D
monitor.
"George" <XXXX@XXXXX.COM>schrieb im Newsbeitrag

"George" <XXXX@XXXXX.COM>writes
your face into said monitor, but I don't think you want to do that.
Cheers,
Ignacio
 

Surprising nobody mentioned this: Why dont you try a 3D hologram ? No glasses/monitors needed and have the ease to set it up anywhere, plus major crowd catcher.
"Ignacio Vazquez" <ivazquezATorioncommunications.com>writes:

it... Repeat the same pattern again and again in x-direction (about 3 or 5
centimeters width). Now each time you want a pixel to appear nearer than the
one next to it, decrease the pattern width by one (simply remove the pixel
from the pattern). Each time you want a pixel to appear farer away than the
one next to it, increase the pattern with by one (insert a new pixel to the
pattern, which can have a random color). Of course you can also increase /
decrease the width by more than one, if you have deep cuts, but smooth
"depth gradients" are best for stereograms. You pattern becomes distorted
the more pixels you have processed, so it is best to start in the center of
the image (with the fresh pattern) and process all pixels to the right,
start again from the center (with the same fresh pattern) and process all
pixels to the left (of course do this for each pixel line). it is hard to
store the pattern (because it can get transformed after each pixel), so it's
more simple to take the pixels you have already calculated instead (which
are a result of the same pattern). One word to the patterns: Using a random
pixel pattern works perfect, using a single-colored (solid) pattern doesn't
work at all, something between works quite well. Most stereogram rendering
tools use a grayscale image for the depths, which makes it easy to measure
the distance from one pixel to the next. If you need some example images to
test your program, look at my homepage at
www.pegtop.de/xfader/stereogram.htm
Jens
P.S. Some pseudo code (I have not tested it, maybe it is completely wrong,
probably you need to add many more 'if this or that')...
for y := 0 to height-1 do begin
pat_width := pattern.width;
for x := center to width-1 do begin
if x-center < pat_width then // take the original pattern
dest[x, y] := pattern[x-center, y mod pattern.height]
else // take the pixels already processed
dest[x, y] := dest[x-pat_width, y mod pattern.height]
dec(pat_width, depth[x+1, y]-depth[x, y]);
end;
for x := center downto 0 do begin
// the same for the left part
[...]
end;
end;
 

"Ignacio Vazquez" <ivazquezATorioncommunications.com>writes
Seriously, if you've got any ideas, I would love to hear about them. You've already
touched the very first problem, I think: identifying things like eg the same tip of
the same nose in different pictures. Going from video, one may assume that subsequent
frames show only slight displacements, and that may perhaps make things somewhat more
feasable...
about that? If there's any kind of tradeoff between extensive but fast algorithm,
possibly packed with a lot of predefined knowledge of the world as it is, and a small
algorithm without predefined structures that takes weeks to compute a 3D mesh of a
room, I think the latter would still be interesting.
I don't just mean interesting from a 'I'm old, but still a child' point of view.
Imagine one could automatically build a 3D world from fantastic views in eg a StarWars
movie. That would be an extremely low-budget way to build a very formidable game
world, for example. And I imagine the '{*word*62}' industry would be very interested to.
The fact that the algorithm does not seem to be already invented may therefore be a
good indication that it is not really feasable. Or is it?
Joris
 

"Joris" <XXXX@XXXXX.COM>writes
3f252f2a$XXXX@XXXXX.COM...
the size of the algorithm, even with the bare minimum of basic assumptions.
And simulating factors like skin elasticity and muscle tone can only make it
larger.
weren't issues, I have little doubt that it would be possible.
Cheers,
Ignacio