Filtering Image Spam Using Image Semantics and Near-Duplicate Detection

Image spam has become the main form of spam, it is a problem crying out for solutions to effectively filter such spam nowadays. This paper proposes an image spam detection system, which is based on image semantics and near-duplicate detection, for solving the problems of current anti-image-spam technologies: low accuracy rate, difficultly recognizing image spam making use of obfuscation techniques and so on. The experimental results show that the system has better filtering effect than previous systems, with increasing more than 10% in accuracy rate and better anti-obfuscation effect, and effectively solves the above-mentioned problems.