Image spam classification based on low-level image features
Chao Wang, Fengli Zhang, Fagen Li, Qiao Liu · 2010
As image spam becomes widespread and does a lot of harm, it is more important to filter such spam effectively for now. In this paper, We propose a feature extraction scheme that focus on low-level features (metadata and visual features) of image, which can making classification rapid. They are effective because of not rely on extracting text and analyzing the content of email. a one-class SVM classifier with RBF kernel as the kernel function is used to detect image spam. Experimental results demonstrate that these features are effective for detecting image spam and comparable to other cutting-age alternatives.