[Spambayes] Inspecting images (was: SpamBayes to HandleEmbedded Images)
Tony Meyer
tameyer at ihug.co.nz
Mon Oct 24 06:42:57 CEST 2005
> Something really needs to be done about this embedded image Spam.
> Honestly,
> SpamBayes appears to be ineffective against all these images,
Can you post an example of a message that is incorrectly classified,
*with the spambayes clues* for the message? The Outlook plug-in
provides this via the "Show Clues for this Message" item in the
SpamBayes menu.
[...]
> I'm sure OCR isn't the only way, but the words are there in plain
> view. It
> seems like the obvious way to resolve this.
Obvious isn't always best. One of the tenets here is "stupid beats
smart" - I think doing some sort of OCR on images would fall into the
"smart" category, and generating simple tokens from the images would
fall into the "stupid" category and be more successful. Just my
opinion, of course, but that's what I'd test if I had time (perhaps
over the (southern hemisphere) summer...or maybe I can convince one
of my employers that this would be worth doing in paid time).
> SpamBayes has been such a great program for me and my colleges,
> family and
> friends. I can only hope that the project sees fit to resolve this
> soon.
It's not really a case of "seeing fit" - the issue is that the
developers are very short on time at the moment (contributions have
always been, and always will be, welcome) and, in addition, this is a
complex problem.
=Tony.Meyer
--
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.
More information about the SpamBayes
mailing list