Prev: how can i transform the binary mask from temps-frequentiel domain to cepstral domain
Next: Tab-expansion freezes matlab, signal for interupting?
From: ImageAnalyst on 4 Jun 2010 16:38
Well I don't know how to do that. Doing a skeletonization on a
megapixel image should be very fast - much less than a second I would
think. The holes problem is no problem - just call imfill (which is
also fast).
From: Gus Lott on 4 Jun 2010 17:29
Cool,

Thanks for the sounding board.

I have something working pretty well, I think. I've got my reconstructed perimeter, the head, and tail, and now I'm chasing counter clock wise from head to tail and taking each point in my fourier smoothed perimeter fit. I then take that point and sweep through all points on the clockwise path around the animal from head to tail and find the point with the smallest euclidean distance.

I then take the midpoint of the line formed by those two points and sweep through all 60 or so points on either side of the animal. there are some caveats near head and tail where the nearest point is on the head, but once you get a bit away from the head and tail, the nearest point is on the other side of the body.

Gives a really nice looking skeleton for this application specific problem and is fairly efficient.

Cheers!
From: ImageAnalyst on 4 Jun 2010 17:36
I would think the nearest other perimeter point would always be the
next adjacent pixel, no matter where it was located.
First  |  Prev  | 
Pages: 1 2
Prev: how can i transform the binary mask from temps-frequentiel domain to cepstral domain
Next: Tab-expansion freezes matlab, signal for interupting?