[ptx] panosifter problems & questions

[Pablo d'Angelo]

On Thu, 12 Feb 2004, scholnik at radar.nrl.navy.mil wrote:

> 
> To add to my earlier whining, I just played around with the "match"
> program that is distributed with the keypoints program.  It has no
> problem finding lots (tens to hundreds) of matches between any of my
> overlapping images with what looks like very few outliers.  What is
> panosifter doing differently that it finds so few matches?  Is there a
> sensitivity parameter to tweak?

The whole panosifter program will be replaced with one that uses
the k-d tree matching. Until then, it'll stay like it is.

the matching algo of panosifter should be very similar to the "match"
program distributed with the keypoints program, but it usually works on
scaled down images.

ciao
  Pablo
-- 
http://wurm.wohnheim.uni-ulm.de/~redman/
Please use PGP