[Spambayes] Spam in Images
Tony Meyer
tameyer at ihug.co.nz
Fri Aug 4 10:23:29 CEST 2006
[Tim Stone]
> I've got a bit of experience with the PIL, if I get a few spare
> minutes,
> maybe I'll mess with something. Can someone send me a typical spam
> image?
Does the recent increase in messages about this indicate that other
people are interested in trying things out? As mentioned previously,
I've been working on these for a while now (when I have the time ;))
and I can put up patches, utilities & result summaries if people are
interested. I've generated tokens indicating the presence & type of
images, sizes (width, height, total), and a variety of things using PIL.
Nothing so far has helped more than hurt, and the closest made the
size of the database balloon. I've been working with my own mail
mostly, since many of the corpora around are old enough that these
messages are pretty rare (I really ought to re-export my mail, I
guess). I also put together a collection of tokenizations as one of
my submissions for TREC2006 (it did ok, but not great, on the public
corpora - still waiting on the private results).
=Tony.Meyer
More information about the SpamBayes
mailing list