Image Spam Filtering by Content Obscuring Detection

Abstract

We address the problem of filtering image spam, a rapidly
\nspreading kind of spam in which the text message is embedded into attached images to defeat spam filtering techniques
\nbased on the analysis of e-mail’s body text. We propose an
\napproach based on low-level image processing techniques to
\ndetect one of the main characterstics of most image spam,
\nnamely the use of content obscuring techniques to defeat
\nOCR tools. A preliminary experimental evaluation of our
\napproach is reported on a personal data set of spam images
\npublicly available.