Tamura Features

26 views
Skip to first unread message

Andrea Ceroni

unread,
May 22, 2015, 11:10:47 AM5/22/15
to jfeat...@googlegroups.com
Hi all,

I'm trying to extract Tamura textures by using the Tamura class. What I get as result is a vector of 18 elements (double), while I was expecting only three values (as specified in the description of the class). Does anyone know why so many values are extracted? And also, what's the meaning of each one of them?

Best,
Andrea

Franz Graf

unread,
May 23, 2015, 5:49:17 AM5/23/15
to jfeat...@googlegroups.com
Hey Andrea,

thanks for using JFeatureLib.
I guess the 18 dimensions are caused by this part of the code in LIRE:
https://github.com/locked-fg/LIRE/blob/master/src/main/java/net/semanticmetadata/lire/imageanalysis/Tamura.java#L277

The array itself is filled a little later in
https://github.com/locked-fg/LIRE/blob/master/src/main/java/net/semanticmetadata/lire/imageanalysis/Tamura.java#L295
So element 0 should be the coarseness, 1 is contrast - followed by the
directionalities.

Hope this helps a little further.

Good luck!
Franz
> --
> You received this message because you are subscribed to the Google
> Groups "JFeatureLib" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to jfeaturelib...@googlegroups.com
> <mailto:jfeaturelib...@googlegroups.com>.
> To post to this group, send email to jfeat...@googlegroups.com
> <mailto:jfeat...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Dr. Franz Graf
http://www.Locked.de
Reply all
Reply to author
Forward
0 new messages