POV-Ray : Newsgroups : povray.binaries.tutorials : 2D to 3D Image Tutorial (500k) : 2D to 3D Image Tutorial (500k) Server Time
28 Mar 2024 11:10:17 EDT (-0400)
  2D to 3D Image Tutorial (500k)  
From: Ken
Date: 5 Mar 1999 07:28:08
Message: <36DFD68A.AF3B4ECA@pacbell.net>
By: Ken Tyler
Date: 03-05-1999


  A simple approach to making 3D objects from 2D images:


   The contents of the enclosed zip file demonstrate how to extract an
 object or part of an image and use it to create a duplicate of that object
 in 3D using a height field as the method employed to give 3 dimensional
 characteristics. There is an assumption that you have the program Leveller
 installed on your system. While not mandatory to the process it can and
 will make it easier for you to make it happen. If you are impaired by the
 OS you use then you can follow the part of the tutorial that explains how
 to extract the image and then use what ever method you need to use to finish
 from there. Probably the single hardest part is getting a good clean image
 to work with and then extractining just the right parts of the image. The
 rest is just proceedural modeling (yeah right).

    I did not plan on releasing this publicly as it was a favour to another
 user. As such it is very rough and not well articulated. If someone finds
 the information useful and learns a new skill then it has more than served
 it's original purpose. I have no plans on developing this any further than
 what you will find with inside the zip file.

   Anyway here is the basics on the included material. I choose a simple
 maple leaf as the object to create. It proceeded quickly in terms of
 processing and took about 4 times as long to document. If you read the
 readme.txt file first it explains what the images in the zip file are
 for. The HF_TUTOR.TXT is the tutorial and there is an example .pov scene
 file provided as well. I suggest you get it out of your system an render
 the .pov file right away. This way by seeing the results possible the
 process part won't sound as bad as it might otherwise.

   If you look carefully at the rendered image from pov there are little
 artefacts visible around where the image border would have been. I missed
 that part when I was blocking out the background and it shows how much
 attention to detail is necessary to get good results. Also I put absolutely
 no extra effort into the finer details that would make this look even better.
 Feathering the edges, cleaner blocking of the background, and a little more
 work spent on some of the large veins in the leaf and this would almost rate
 as a classic pov object.

   The original .jpg is included for reference and I'm sure you will agree
 the process has more possibilities than many people would think. When you
 look at the original image you will see that it is almost trash file quality.
 If you can take an image this poor in quality and produce something useful
 with it just imagine what you can do with a high quality image to start with.

  Good luck and if I can clear up any confusion or answer any questions about
 the process I'll be around. Just give a yell. Jim Kress has a good feel for
 the process now and can also answer specific questions related to the project
 he just finished.


-- 
Ken Tyler

mailto://tylereng@pacbell.net


Post a reply to this message


Attachments:
Download 'hf_tutor.zip' (427 KB)

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.