Execution time of classification (Q4a)

Moderator: Computer Vision

DanielR
Mausschubser
Mausschubser
Beiträge: 83
Registriert: 19. Feb 2008 13:15

Execution time of classification (Q4a)

Beitrag von DanielR »

Hi there.

Is anybody also observating the same time for the different classifications methods? We only have small deviations that I expect to be caused through threading and other GDI3 stuff. ;)

On my maschine, for example, we get for the three methods:
[19.9731, 19.6655, 19.8223]
when we consider training and classification.

I'm confused about that, because I thought the question was aiming at efficiency of different mathematical operations in matlab. But I've talked to one other group that is essentially experiencing the same thing.

Christoph Vogel
Erstie
Erstie
Beiträge: 20
Registriert: 30. Mär 2010 17:38

Re: Execution time of classification (Q4a)

Beitrag von Christoph Vogel »

This is probably because your evaluation is currently dominated by the computation of the descriptor.
The question was aiming at the evaluation of the SVM and not at the whole process of descriptor extraction.

But I think that is not completely clear from the task description.
So you could just mention that extracting the descriptor is the hard task here.

Best regards,
Christoph Vogel

DanielR
Mausschubser
Mausschubser
Beiträge: 83
Registriert: 19. Feb 2008 13:15

Re: Execution time of classification (Q4a)

Beitrag von DanielR »

Thanks for the answer.

As we just "timed" the given learning and classification methods and these also contain the descriptor calculation , we really include the descriptor extraction in our measurement.

Antworten

Zurück zu „Computer Vision“