I have been experimenting lately to determine some of the aspects which make for a good face synth.
This last set of photos of Mercedes was done in regular room lighting and my lighting exposure was way off. But it still turned out not to bad.
My plan now is to try to create a (close to) photo realistic human face synth, with natural lighting in a natural setting, using my photographic methods.
If I can do it with Mercedes, I am pretty sure I can do it with a human face.
The only "editing" done on this set of images, was to delete the blurry images.
The only time I will use my post processing methods, will be to enhance weak point cloud areas.I will rely strictly on original images photographed using one or more camera orientation constants.
So far I have used only one constant in any of these experimental face synth
and very limited post processing. 
http://photosynth.net/d3d/photosynth.aspx?cid=1b25764d-fb64-4167-a9ca-ce2ecaf9c3c

Looks like the last character of your collection ID somehow got deleted.

Here's the link: 
http://photosynth.net/d3d/photosynth.aspx?cid=1b25764d-fb64-4167-a9ca-ce2ecaf9c3c3

I am trying not to commit to any expectation of the outcome of the planned face synth, but if all goes well, I hope this will demonstrate the effectiveness of my photographic methods without the need to use any of my post processing. If not, oh well...nothing ventured, nothing gained. The hardest part of a human face synth, is to find someone willing to have their face photographed for all to see.
My older brother bravely offered his face for earlier tests, and has agreed to be my subject. I am pretty sure that I would not volunteer as he has, so I am very  grateful for what he has agreed to do for me. 
Stay tuned and wish me luck.

Thank you Nathanael.
What time is it where you are?

It is 2:34 a.m. here and I am off to bed.