QuickTopic free message boards logo
Skip to Messages

TOPIC:

Pictorial Structures for Object Recognition

  Messages 6-5 deleted by author between 07-22-2006 02:06 AM and 07-21-2006 09:00 AM
4
Robin Hewitt
11-16-2004
11:17 AM ET (US)
I was curious about the reasoning behind some of their choices. In particular, why the use of set (rather than centroid) distances, and why the choice of mahalanobis (rather than euclidean) metric for measuring grid displacements.
Edited 11-16-2004 11:18 AM
3
Rasit Topaloglu
11-16-2004
03:33 AM ET (US)
I have seen that spring-based models work quite well for a wide number of engineering applications

The detection of occluded images is important to prove for this type of a model (and is done in the paper experimentally). But they should also have done occluded images of people. Say we only see a person's body and head, it would be quite possible to confuse the person with other objects (like TV on a table). Some of these can be eliminated using the prior probabilities though by restricting the size of the persons.
Edited 11-16-2004 03:46 AM
2
Stephen Krotosky
11-16-2004
02:27 AM ET (US)
Just the parameters. For example, in the face detection case, they decided to use derivative of gaussian filters and the training set's response to these filters is used to learn the model parameters.
1
Hamed Masnadi-Shirazi
11-16-2004
01:36 AM ET (US)
Is the Model Structure learned as well or just the parameters of the imposed model structure?

Print | RSS Views: 1932 (Unique: 922 ) / Subscribers: 0 | What's this?