|
3D to 2D Conversion
 |
Do any one of you 3D veterans know of a relatively simple way of producing a depth map from a stereo pair?
I'm thinking something in the line of simple PS actions, plug-ins, or the like.
3D rendering applications and raster to vector conversion would be beyond my reach.
A depth map comes in extremely handy for trimming anaglyphs, or merging two anaglyphs together.
Posted at 8:56PM, 3 August 2008 PST
(
permalink
)
|
 |
You can use various forms of selecting and save them as paths. You can fill these with shades of gray. Unfortunately the object outline has to be horizontally wider in your dephtmap than in the image. Therefore all areas has to be expanded horizontally according to how far from RGB 128 they are, and therefore the answer is no. It´s proberly just a programmers issue and a company called DDD actually had a tradeshow here and they claimed their software could generate dephtmaps from stereopairs, but their operators were not able to make it work when they had the chance. The problem is pretty simple, but the solution might not be. I imagine they are still working on it.
You are right jay, the dephtmap is perfect for trimming. In fact it is most efficient and fantastic tool for anaglyphers ever made.
Posted 4 months ago.
(
permalink
)
|
|
Thank you Louis for your suggestions. What I'm trying to achieve is to transform a selection (path?) made in red channel, to the corresponding area in the green or blue channels. I wonder if depthmaps work on paths?
Posted 4 months ago.
(
permalink
)
|
 |
Not sure I fully understand what you want to do, but could you split the channels to red, green and blue, work on them and then recombine? (Maybe there is a way to work with the channels as layers once you've split them).
Posted 4 months ago.
(
permalink
)
|
|
Dan, working on a stack of layers is given, channels can be used as extra layers to help with making selections. If the image contains well defined features, it would be easy to trace these features in each channel. The difficulty arises when the picture elements are not adequate to allow "connecting the dots".
The problem is that a selection made in one channel can not be directly superimposed on the matching channel for the other eye, due to the parallax. The idea is to automate the transformation necessary to match the selections. A depthmap does the transformation on bit mapped images. We're trying to skip the manual construction of a depthmap when the depth information is already available in the form of parallax between the two views.
Originally posted 4 months ago.
(
permalink
)
jaysdesk edited this topic 4 months ago.
|
|
To my knowledge no jay. But you can displace each channel with a dephtmap.
Posted 4 months ago.
(
permalink
)
|
Would you like to comment?
Sign up for a free account, or sign in (if you're already a member).
|
|
Anaglyph (Red-Cyan 3D Pictures) /
Discuss
Feed – Subscribe to Anaglyph (Red-Cyan 3D Pictures) discussion threads
Flickr Blog |
About Flickr |
Terms of Use |
Your Privacy |
Copyright/IP Policy | Report Abuse
|
Deutsch
|
English
|
Español
|
Français
|
|
Italiano
|
Português
Copyright © 2008 Yahoo! Inc. All rights reserved.
Bookmark on Delicious