Re: [CinePaint] Looking for a feature useful to smooth timelapse sequences - grading

On 11/22/06, Bob Friesenhahn <[email protected]> wrote:
> > As much trouble as I'm having just communicating what it is I'm trying
> > to do, can you see why I am having trouble searching for a tool that
> > does it?  I don't know, maybe you would call it inter-frame exposure
> > or levels smoothing or something like that...
>
> It seems like what you need is a tool that analyzes the image minimum,
> maximum, and mean, for each image over a sequence of images. Then it
> would compute a correction value for each image.

That's more or less what I am looking for.  I created my own, but the
correction values that I created weren't right.  :(

> Maybe it would
> smooth the computed values using a filter function (across frames) in
> order to allow gradual changes over time.  ImageMagick's -level option
> can then be used to adjust each image's parameters.  The -level option
> works similar to Photoshop's "Adjust Levels" dialog.

Yes, that is exactly what I intended to use.  Levels, or curves, or
whatever it takes to make it work nicely.

> This sort of work is likely very time consuming and the results are
> dependent on the nature of the images and what is intended.

Time-consuming, I can live with, if it is 100% automated.  Once I have
something that WORKS, I can work on making it work faster...

Realistically, the ability to correct one image to make it match as
closely as possible the other image is what I need.  I can loop over
the image, pixel by pixel, and compare.  What I'm not sure of is what
to do with those results?

Problem is, the min/max of the frame image may not be the same as that
in the target image.  Because the average has things 'smoothed' out,
some (moving) sharp edges may be fuzzy.  Thus, a simplistic approach
of 'normalization', even if it is just fitting it to the histogram of
the average image, isn't going to work.

Now, for a given pixel, there may be a very large number of possible
levels or curves adjustments that could be done that would cause the
current pixel to match the average pixel.  I could just take the one
that works for the most pixels.  But that seems like a very
brute-force method.  There's got to be a better way...

-- 
--------------------------------------------------
John Arrowwood
John (at) Arrowwood Photography (dot) com

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV