Assuming I start with only a digital camera (no photo editing software
installed)...

....how would I end up with an outcome that stacks three headshots of
different people on top of one another....blend/edit them
together...and wind up with "one headshot" that is a representation of
how the three people would look as one combined "face"?

Please include the format I should use for the output of the original
three digital photos, some software titles of the required editing
applications, and as much of a "hold my hand" step by step tutorial
for how to do this.

Assume this is being written for a beginner...because that's what I am.  :) :)

---

Request for Question Clarification by landog-ga on 02 Aug 2005 09:58 PDT

Hi
Try this freewware program:

Fanta Morph 1.0 

FantaMorph! is a powerful and easy-to-use morphing software for the
creation of fantastic image morphing pictures and sophisticated
animation effects. With our revolutionary rendering engine and a super
friendly user interface, creating morphing pictures and action
sequences from still images has never been easier and faster!

Play final morphing effects in real time without exporting to a file 
Instanty morph one image into another!, cool animations! 
Add unique pictures & animations to your web site, wallpaper,
advertising, presentations, training materials, and cards.
Create pictures with shadows, reflection, and perspective 
Animations with sound using multiple, textured 3D models on
backgrounds with 3D effects
Animations with moving backgrounds and motion blur effects. 1,100+
models for users to download. Features: import function for VRML
models
3D rotate, pan, and zoom 
Shaded or polygon display 
User defined transparency, colors, textures, motions, orientations,
light position, and sounds
3D model editing 
3D effects for 2D backgrounds 
4 user selectable picture quality levels 
user definition of depth & orientation for 3D models at arbitrary
points on the animation path
special motion options including closed paths, circular paths, and rotations 
depth specification for elements of backgrounds 
functions for merging up to 30 model motions into realistic 3D movies
with user-defined frame counts
Table driven motion and sound editing.

Download:
http://www.gold-software.com/FantaMorph-file3990.html

It can do some nice blending.

---

Clarification of Question by submaniac-ga on 02 Aug 2005 12:25 PDT

landog....thank you for the start in the right direction. This is the
email I sent to FANTAMORPH (both their sales and their tech support
folks)...I will follow up with their response when it arrives.Any
additional insight regarding the original question is greatly
appreciated!

EMAIL SENT:

Hi there....I was referred to your product and your site...and just
have one specific prepurchase question.

  

Assuming I start with only a digital camera (no photo editing software

installed)...

 

....how would I end up with an outcome that stacks 50 headshots of
different people on top of one another....blend/edit them
together...and wind up with "one headshot" that is a representation of
how the three people would look as one combined "face"?

 

To be clear...I'm not as interested in producing the ?morphing effect?
that makes it look cool as it goes from one image to the next. What I
want the final outcome to be is one ?picture? that has blended all of
the faces from the 50 individual photos and shows one face with the
combined characteristics of the underlying photos.

 

Can this be done?

There is not likely to be much out there to assist you.  What you are
asking for, essentially, is a multi-source morph.  I.e. if one nose is
fat, and one nose is thin, and one nose is average but bent, you would
combine the shape of all 3 and average them out, hopefully ending up
with an average-sized nose with a slight bend in it (as 1 out of 3 had
the bending).

Simply "blending" the photographs will not work, as this will create
obvious, multiple frames.  Also, the position of the headshots would
be required to be compeltely identical for all of the headshots, right
down to the level of zoom.  In the above scenario, "blending" the
photographs would produce a picture where you would see one general
face shape but three, each-slightly-transparent noses.

A morph is traditionally used to combine the features from two
sources.  A morph is typically done by creating reference points on
two different pictures that are the same: for example, a reference
point of the outside corner of the left eye.  By mapping the same
reference point to each picture, and then repeating this process for
the entire "detail" of both photographs, a computer program can blend
the two pictures and "warp" them so that the reference points end up
at a happy medium.

Theoretically, it should be possible to blend multiple photographs as
long as they all had the same reference points.

That being said, it will be impossible to combine all of the features
of a large group of people into one face.  If twenty-three people are
very fair skinned and four people are very dark skinned, the
"combined" montage will look very fair skinned, and you will have
captured very little of the essence of the darker skinned people.  The
same would go for a large nose, a pointy ear, blue eyes, or whatever. 
The more people are in the photo, the more obvious this "averaging"
will be.  (If you've ever superimposed colourful photographs, you
learn very quickly that combining multiple, complex things just yields
bland results).

I hope these comments are helpful to you.

ryanjblack...thank you for telling it like it is.

Drat! :(

"ryanjblack...thank you for telling it like it is."

Well, I told it like it is... in my opinion ;)  There may be some
digital photography expert out there who knows better than me!  I'd
recommend leaving the question open for a while longer and seeing if
anyone else has any suggestions.

I've given this some more shot.

If you had a small number very similar photographs and access to a
free morphing program, here is how you could combine the features of
the photographs.

Step 1: import first 2 photos from the stack into morphing program
using photo1 as a source and photo2 as a source.

Step 2: outline the reference points on each photo (left corner of
eye, top left side of eye, top of eye, top right side of eye, right
corner of eye ... lol there's lots of reference points, this is a
looooong step)

Step 3: have the morphing program produce the 50% result (i.e., the
morph that is 50% between photo 1 and photo 2)

Step 4: output that result to a new photograph (this is what i call
the "morphed stack")

Step 5: import the resulting morphed photograph (the "morphed stack")
and the 3rd photo from the stack into the morphing program, using the
morphed stack as a source and the 3rd photo as a source.

Step 6: if the reference points are preserved from the previous morph,
then you only have to create the same reference points for the newly
added photo.  Otherwise, redo the reference points (Again, a very long
step)

Step 7: have the morphing program produce the 66% result (i.e., the
morph that results when 66% of the morphed stack and 33% of the 3rd
photo from the stack is used)

Step 5: import the "morphed stack" and the 4th photo from the stack
into the morphing program, using the morphed stack as a source and the
4th photo as a source.

Step 9: see step 6

Step 10: have the morphing program produce the 75% result (i.e., the
morph that results when 75% of the morphed stack and 25% of the 4th
photo from the stack is used).

You can repeat this as many times as you like.  Each time, increase
the ratio of the morphed result and decrease the ratio of the photo
you are adding to the "morphed stack".

Here's how the ratio would work:
2 photos:  1/2 vs 1/2
3 photos:  2/3 vs 1/3
4 photos:  3/4 vs 1/4
5 photos:  4/5 vs 1/5  (i hope you see the patern here ;D)

It would be tedious, but with a small stack of photos, a digital
camera, and a freeware morphing program, it could likely be done.

Cheers,
Ryan Black

ryanjblack...thank you again, sir!

Sounds like a scenario where juuuuuuust about the time I have morphed
in the 500th photo...the new version of software will be released that
does it all in one step. Lol!

Hmmm....I think I just embraced a new phrase for the day.

"low cost, outsourced labor".

:) :)

All good ideas.

One other method I think is interesting to play around with is
Photoshop software or similar. You can layer stacks and stacks of
headshots and also tweak each layer for transparacy and add a
gazzilion effects to each layer etc etc. Maybe you'll be creating
something reallly cool.

Success!

Using Photoshop: You can make each head-shot a layer as suggested, but
apply a mask to each image, the mask button is on the bottom of the
layers panel, the second one that looks like a box with a with dot in
it, you can then "paint" (using the brush) with black and white within
the mask to reveal or hide parts of the image, if you press alt when
you create the mask(press the mask button) it will fill the mask with
black hiding the layer, it normally starts off white. You then have
much more control because you can keep the original layer intact and
use the brush at what ever opacity you need in order to bring the part
of the image that you want in or out. Only use black and white when
painting and make sure you are in the mask portion of the layer, it
will show a box next to the thumbnail image showing where you have
painted on each layer.

i have a degree in Photography and digital Imaging and have not heard
of something that will "do this for you" im not sure if you were
looking to choose the pieces of the faces your self or have a program
create it but this is the most nondestructive way i can think of to
achieve what you are looking for