Jump to content

Blue screen matte COP VOPNET


edward

Recommended Posts

I'm putting this here hoping it will get moved to the compositing forum as I can't seem to upload files there.

For the topic, please see:

http://odforce.net/forum/index.php?showtopic=2339

---------

I implemented a COP VOPnet a while back based on the ultimatte equations from Jim Blinn and Alvy Ray Smith's paper, "Blue Screen Matting" (ftp://ftp.alvyray.com/Acrobat/BluSig96.pdf) which actually worked pretty well for the test images I tried.

---------

If you look at the VOP network, the Parameter VOPs are named according to the paper. The workflow for using it goes something like thise:

- Start off with the default parameters.

- Look at the A plane of the bluescreenkey cop

- Adjust each parameter in turn except for the first and last 3 until you get as good a looking matte as you can. Typically, adjust it so that you get the areas you want as white as you can and the areas you don't want as dark as you can.

- Once you're done that, adjust the Lower/Upper Alpha clamp parameters to isolate the range of the alpha you want to clamp at. Typically, you just need to raise the Lower Alpha clamp value high enough so that it completely remove the blue screen.

- Look at your result and then try tweaking the parameters again. Typically, I haven't found the Global Adjust parameter useful.

- The last parameter can be used to add back in some blue if the matte removed too much of it.

bluematte.zip

Link to comment
Share on other sites

Yeah, I know. :( The problem is then I need to clean it up, add comments, get an account on the exchange, etc... :) ... Plus, I never really played with it long enough past my test images (of at most 3). Besides, there's way more interesting papers on doing matting these days (like graphcut, bayesian, optical flow, etc) that it seems hardly worth the effort on something so primitive. :)

Link to comment
Share on other sites

Yeah, I know. :( The problem is then I need to clean it up, add comments, get an account on the exchange, etc... :) ... Plus, I never really played with it long enough past my test images (of at most 3). Besides, there's way more interesting papers on doing matting these days (like graphcut, bayesian, optical flow, etc) that it seems hardly worth the effort on something so primitive. :)

14997[/snapback]

You can log in with your (SESI) forum information.

Link to comment
Share on other sites

  • 9 months later...
  • 2 months later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...