Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > Auto-mirror an image?

Reply
Thread Tools

Auto-mirror an image?

 
 
Armando
Guest
Posts: n/a
 
      10-25-2004
Hi all -

First time poster here (for a great while anyway) - I've lurked and not seen
this discussed, hope that lets me in...

I had long ago heard that if you mirror an image of a person's face (two
left or two right halves together), you can see two different
"personalities" emerge. I tried it myself, and immediately realized that
the "personalities" are due mainly to side-lighting, and maybe some facial
asymmetries. With further playing, I started to get some pretty hilarious
results, especially by putting the "fold line" off-center or skewed on the
face.

While I've gotten pretty adept at the mouse- and keystrokes required to whip
one of these up quickly by now, I'd really like a way to do it "live", where
I can actually view the image I'd get given where the camera is currently
pointing. I'm not sure what kind of output the camera would need (my Sony
DSC-S50 has a live TV out, but that's analog/raster/whatever and maybe not
right). I assume the conversion device would have to be hardware as opposed
to software, but I know nothing of what is out there.

Can anyone give me some ideas as to what I'm probably going to need to look
for to do this?

Thanks one and all,

Armando


 
Reply With Quote
 
 
 
 
Gene Palmiter
Guest
Posts: n/a
 
      10-25-2004
Can't this be done without all the technology? Just use a mirror and take a
photo of it.

"Armando" <(E-Mail Removed)> wrote in message
news:_0_ed.8040$(E-Mail Removed)...
> Hi all -
>
> First time poster here (for a great while anyway) - I've lurked and not

seen
> this discussed, hope that lets me in...
>
> I had long ago heard that if you mirror an image of a person's face (two
> left or two right halves together), you can see two different
> "personalities" emerge. I tried it myself, and immediately realized that
> the "personalities" are due mainly to side-lighting, and maybe some facial
> asymmetries. With further playing, I started to get some pretty hilarious
> results, especially by putting the "fold line" off-center or skewed on the
> face.
>
> While I've gotten pretty adept at the mouse- and keystrokes required to

whip
> one of these up quickly by now, I'd really like a way to do it "live",

where
> I can actually view the image I'd get given where the camera is currently
> pointing. I'm not sure what kind of output the camera would need (my Sony
> DSC-S50 has a live TV out, but that's analog/raster/whatever and maybe not
> right). I assume the conversion device would have to be hardware as

opposed
> to software, but I know nothing of what is out there.
>
> Can anyone give me some ideas as to what I'm probably going to need to

look
> for to do this?
>
> Thanks one and all,
>
> Armando
>
>



 
Reply With Quote
 
 
 
 
Dave Martindale
Guest
Posts: n/a
 
      10-25-2004
"Armando" <(E-Mail Removed)> writes:

>While I've gotten pretty adept at the mouse- and keystrokes required to whip
>one of these up quickly by now, I'd really like a way to do it "live", where
>I can actually view the image I'd get given where the camera is currently
>pointing. I'm not sure what kind of output the camera would need (my Sony
>DSC-S50 has a live TV out, but that's analog/raster/whatever and maybe not
>right). I assume the conversion device would have to be hardware as opposed
>to software, but I know nothing of what is out there.


>Can anyone give me some ideas as to what I'm probably going to need to look
>for to do this?


If you're a programmer, this may be possible to do in essentially real
time using current graphics cards. The idea would be to read in a
video image and store it in memory every 1/30 second (obviously, you
need a video card with video input). Then tell your graphics card to
draw two rectangles, one filling the left half of the monitor and one
the right half of the monitor. Use the current video image as a texture
to texture-map the two rectangles. Now, by setting the (u,v) texture
coordinates of each of the corners of the rectangles, you can determine
exactly what pixels of the texture get mapped to the rectangles on the
screen, so you can do the mirror-imaging easily - as well as
magnification, rotation, and lots of other fun things. The texture
mapping hardware of the graphics card will take care of all of the
address calculations, image filtering, and so on.

Dave
 
Reply With Quote
 
Armando
Guest
Posts: n/a
 
      10-25-2004
An obvious but low-tech solution, thanks. What I'm after is a seamless
transition at the fold, and full quality on both sides. I imagine a kid's
kaleidoscope with a folded sheet of metal as the mirror pair - looking in, I
can easily see which section is the actual, and which sections are the
reflections. Using a first-surface mirror would probably come close, but I
don't want a klunky (and fragile) apparatus on the camera. I don't mind
doing some coding if necessary, or spending a bit on hardware (both of which
might let me create additional effects as I play), but I don't think the
answer is optical.

Thanks for the reply, sometimes the obvious IS the right answer but gets
overlooked, so it never hurts to suggest it.

Armando


 
Reply With Quote
 
Armando
Guest
Posts: n/a
 
      10-25-2004
Dave -

I can't remember if it's top- or bottom-posting that ****es people off, so I
won't include your reply...

This is getting closer to what I think I'll need to do. I don't mind the
programming, it might be a challenge but then I'd be learning something new,
which is never bad (especially since I'm out of hi-tech for a bit too long
now).

Your solution sounds feasible, but I suspect the video out from the camera
is lots lower resolution than what the CCD can deliver - is there another
way (or another type of camera) that will give me VGA (or hopefully better,
up to say 1600x1200) output in real time? How would that get into the PC in
a way I can grab it?

Thanks again for your thoughts!

Armando


 
Reply With Quote
 
Dave Martindale
Guest
Posts: n/a
 
      10-26-2004
"Armando" <(E-Mail Removed)> writes:

>Your solution sounds feasible, but I suspect the video out from the camera
>is lots lower resolution than what the CCD can deliver - is there another
>way (or another type of camera) that will give me VGA (or hopefully better,
>up to say 1600x1200) output in real time? How would that get into the PC in
>a way I can grab it?


Video out from a digicam may or may not be full video resolution, but
video output from a good video camera or camcorder should be equivalent
to 640x480 pixels at 30 FPS (higher res, lower frame rate in the case of
PAL). Standard graphics boards with video in will handle this.

If you really need higher resolution, you'll need a HDTV video camera
and special video input hardware.

There are digital still cameras that will give you megapixel images
under control of a computer, but you're now talking about seconds per
single image, not 30 images per second.

Dave
 
Reply With Quote
 
Armando
Guest
Posts: n/a
 
      10-26-2004
Dave,

Thanks again for the return info. All I really want in "real time" is the
ability to see what the camera is pointed at, with the mirror (or other neat
effect) applied. That can easily afford to be less than 30 FPS, since all I
need to do is compose and then grab a frame when I like it.

We can take this to email if it starts to have less interest to other
readers:
armando_odie (at) att (dot) net

Armando



 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off




Advertisments