Research on Content Based Image Spam Filtering Technology
Yanyan Liu · Journal of Liaoning University of Technology · 2014
The image-based spam filtering scheme was proposed, which does not depend on the text label of the image, however the visual feature is extracted directly, including the image gradient histogram, color histogram and LBP features in the scheme. The Support Vector Machine algorithm was analyzed. Based on this algorithm, the image spam filtering was implemented. The experimental results show that LBP is the best as compared with the image gradient histogram and color histogram.