Capturing real heads and faces for modelling

I want to create models and renderings of actual people that I know. Anyone have experience or advice on software, hardware, etc. They don’t have to be such high quality that you couldn’t tell the renderings from the real thing, but where someone would say - “hey that looks just like…”

You can do pretty well with just a camera. Take some good shots to use for reference, and for texture mapping. Front view, side view, 45* view, etc. If the subjects are willing, an interesting technique is to draw a grid/edge loops on the person’s face with a marker pen so you can trace the contour lines with 3D curves to get a 3d shape that accurately mimics the face shape.

Take a look at the ‘Subsurf Head Modeling’ thread by Luckybreak, especially the tut-link posted by Kos.


hey…wake up…the tute link i mentioned has been created by me!! 8)
btw… while capturing images of faces for reference i think you should keep the zoom of the camera as high as possible i.e. 70mm or more.i don’t have experience with capturing faces but i think keeping the zoom up will reduce the perspective errors… and thus the errors in your model too.and also when the head is finished try putting the camera in blender or whatever software you use in such a way that you can compare the 3d-head with the reference image in the bacground of the camera.errors can be corrected in this is a experimental process though and i have not tried it seriously yet.some day you may see a tute for that.

Try PhotomodelerLite

you can still get it at

I just did the tutorials and then sucessfully imported the VRML of the Blocks tutorial into Blender. The UV mapping and all came in great.

A few minor tricks when importing.

  1. The VRML object came in extremely small
  2. The faces did not render until I assigned a material with Texface pressed
  3. It would not render properly, then I toggled edit mode and draw normals and everything showed up.
  4. Faces from each photo came in as a separate objects. Join and remove doubles if you want to subsurface - otherwise you get holes.

I would love to see someone try this with a head. I will give it ashot when I have time.

Borken and Kos are right, the better the reference pictures are the easier the modeling.
I once found a front, side, and back :o view image of a sculpture head. Minor differences in the horizontal camera angle made the modeling a nightmare. Sometimes this confused more than the references pics helped.
To compensate perspective errors a bit you can try non uniform scaling in PS or a similar app.

(An alternative would be a 3D scanner. Minolta just offers one for only about 20 000 bugs (+software). :o :o :-? :o :-? :o )

Kos wrote:

hey…wake up…the tute link i mentioned has been created by me!!