a powerful software: Human Generator

http://www.shusaku.co.jp/www/product06_e.html
because there is no software available, I put the information on
“News & Chat” board other than “Other software”

     We are just looking for reseller in the US and Europe. It is not on sale now. Please wait until release.

     Take a portrait from front side and trace its contour
     Input feature points that as eyes, mouth, nee, and so on. 
     Then trace contour of the picture. Within 5 minutes (includes operation time), you can finish.
     Automatically, 3D model is generated!
     HumanGenerator generates any types of body shapes automatically. For example, you can make both thin people and fat people. And HumanGenerator generates many hairstyles and many types of cloth.  
    Change pose! You can change human pose from "pose library".

http://www.shusaku.co.jp/www/images/Woman3D.JPG

uhhh, makehuman? :smiley:

Looks like a scam to me. This sort of stuff is just not possible at the moment.

Well, it is actually with varying degrees of accuracy. You can input photos of yourself into Tony Hawk’s pro skater and choose facial features - this is just a little more sophistimacated.

What you really want is what they do in the Matrix with high-res photography on agent smith and then they can do photoreal stuff:

The universal capture system uses an array of synchronized cameras to capture an actor’s performance in ambient lighting. Each pixel’s motion is tracked over time in each camera view using optical flow and then combined with a cyberscan model of a neutral expression of the actor and photogrammetric reconstruction of the camera positions. Animated color maps for the model are created by combining images taken from the multiple camera views over time, and surface textures such as pores and wrinkles are added using a 100-micron scan of the actor’s face taken with an Arius3D scanner from which bump and displacement maps are extracted.

But for cloth physics and a model setup like that in 5 minutes it looks ok. It better be free though. I’m not payin’ for stuff like that when you can just use a modelmaker like makehuman as someone mentioned.

Sticky texture just wouldn’t do.
But autotrace 2d images might be handy for makehuman.

Well, it is actually with varying degrees of accuracy. You can input photos of yourself into Tony Hawk’s pro skater and choose facial features - this is just a little more sophistimacated.

What you really want is what they do in the Matrix with high-res photography on agent smith and then they can do photoreal stuff:

The universal capture system uses an array of synchronized cameras to capture an actor’s performance in ambient lighting. Each pixel’s motion is tracked over time in each camera view using optical flow and then combined with a cyberscan model of a neutral expression of the actor and photogrammetric reconstruction of the camera positions. Animated color maps for the model are created by combining images taken from the multiple camera views over time, and surface textures such as pores and wrinkles are added using a 100-micron scan of the actor’s face taken with an Arius3D scanner from which bump and displacement maps are extracted.

But for cloth physics and a model setup like that in 5 minutes it looks ok. It better be free though. I’m not payin’ for stuff like that when you can just use a modelmaker like makehuman as someone mentioned.[/quote]

It’s impossible to make a relatively accurate 3D model from a static 2D image. You’d pay at least AU$50 000 for a 3D scanner that can get a slightly better result.

And if you look at their website you can see how “Proffesional” it is.

Sorry if you were joking osxrules, but it kinda sounded like you were being serious. %|

Not everybody needs spiderman ‘profesional’ solutions.
sometimes it’s handy to have some different persons as image fill.
Make human or poser are excellent to create such models.
But I can imagine that instead of typing numbers or pulling sliders, tracing 2d images can be a good alternative to set the variables in thoose makehuman scripts. a ‘director’ chooses the kind of people it wants, the intern traces the scans. No need to know how ‘make human’ or poser works, just how to trace an image. So yes I think osxrules takes the idea seriuos, so am I.

But notice how you don’t see any of these models from the back :slight_smile:

Yes, you see it! just a little bit, on the girl under “back hair” section… and you can see that back is like front!!! (see the white shape of the shirt…)

ah
probably keeps the colors from the edges.