A novel anti-spam scheme for image-based email

Spam has become a serious problem not only to the Internet but also to society. In past years, researchers studied the problem mainly on textual features. As the content of spam removing from text-only to multimedia-enriched, the existing text-based spam filters are no longer effective for these emails. This paper proposes a novel filter approach that focuses on the spam which contains images. The approach based on detection and counting the similar image-based emails. A special challenge-response mechanism also is introduced in this paper in order to avoid legitimate being chucked. By this mechanism, users no longer need cost time to check the spam folder.