Ex 2 1b)

Moderator: Computer Vision

Marc Schmitt
Neuling
Neuling
Beiträge: 7
Registriert: 29. Apr 2010 12:12

Ex 2 1b)

Beitrag von Marc Schmitt »

Hi @all,

we have a question on the exercise mentioned above.

In this thread it was said by Diane, that all histograms should be normalized. So, we wonder which normalization method should be used.

We first think that we should normalize using the MAX-norm, but then look at the m-file for the Dx/Dy-histogram. In this file it was said to normalize the histogram, so the integral of it is equal 1.

So, to come to an end ;), should we use different normalizations for the different histograms, or should we normalize all histograms using the integral norm?

Thanks for your help.

Benutzeravatar
leviathan
Computerversteher
Computerversteher
Beiträge: 307
Registriert: 30. Jul 2008 14:26
Wohnort: Darmstadt
Kontaktdaten:

Re: Ex 2 1b)

Beitrag von leviathan »

I would suggest using the integral norm for normalizing all hystograms, because the comparison functions (intersect, for instance) expect them to be normalized with it, and they can all be used in combinations in the matcher function.
Ein Programmierer hat immer eine Lösung. Die passt nur nicht immer zum Problem.

Hiwi für Weiterentwicklung des Lernportals (Moodle).

Antworten

Zurück zu „Computer Vision“