[Spambayes] date for new release to handle image spam?
skip at pobox.com
skip at pobox.com
Fri Feb 2 16:27:06 CET 2007
Seth> Rather than try to imagine which clues will be definitive, I was
Seth> thinking out loud if we might provide a large number of seemingly
Seth> unrelated clues and letting the Bayesian classifier look for
Seth> correlations.
Yeah, but we do need to perform the clue extraction from the image or its
properties. ;-)
Seth> Maybe things like animation rate, contrast ratio, color bias,
Seth> ... any actual piece of information that varies from one image to
Seth> the next.
If people can tell me how to compute any of these metrics using PIL (or
point me to some cookbook sites that describe them), I can put them into SB.
Seth> Then there are the email specific ones like content transfer
Seth> encoding of each MIME part, total characters in each MIME part,
Seth> character set, etc.
I think a fair amount of this stuff is already calculated. We probably need
a dictionary of synthetic clues written down somewhere we can refer to.
Skip
More information about the SpamBayes
mailing list