QuickTopic free message boards logo
Skip to Messages

TOPIC:

Graphical Models for Graph Matching

5
Paiton
07-21-2006
11:31 PM ET (US)
Nice work! Comprehensive discussion. Visit my page too: cipro xl webpage devoted to cipro xl. atenolol and chlorthalidone webpage devoted to atenolol and chlorthalidone. I would just like to congratulate you on doing an amazing job!
4
Deleted by topic administrator 07-21-2006 09:00 AM
3
Stephen Krotosky
10-26-2004
04:03 PM ET (US)
It's interesting that they show only one example result and only claim feasibility, not generality in practice. I wonder how much tweaking they needed to get good results.
2
Robin Hewitt
10-26-2004
01:51 AM ET (US)
Rasit - yes, that must be right: p(y_i|x_i=s_alfa) instead of p(y_i|x_i=x_alfa). Nice catch!

MAP = maximum a posteriori (probability).

Maybe someone else will want to answer the question about neighbors, so I'll hold off on that....
1
Rasit Topaloglu
10-25-2004
06:35 PM ET (US)
The Markov Field is supposed to be created using Gx. However, how do we choose neighboring nodes for the Markov model? (I think that the number of nodes in the Markov model is the same as the number of nodes in Gx. But a node in Gx may have more than one neighbor.)

Correct me if I'm wrong but I think on page 3, it should be p(y_i|x_i=s_alfa) instead of p(y_i|x_i=x_alfa)

What is a MAP criterion, mentioned on page 3? Is it a well known name for Equation (1)? Overall, this paper defines unknowns from bottom to top, it is so distracting.
Edited 10-25-2004 06:56 PM

Print | RSS Views: 2096 (Unique: 1050 ) / Subscribers: 1 | What's this?