Image spam classification using neural network

Abstract

Spam, an unsolicited or unwanted email, has traditionally been and continues to be one of the most challenging problems for cyber security. Imagebased spam or image spam is a recent trick developed by the spammers which embeds malicious image with the text message in a binary format. Spammers use image based spamming with the intention of escaping the text based spam filters. On the way to detect image spam, several techniques have been developed. However, these techniques are vulnerable to most recent image spam and exhibit lack of competence. With a view to diminish the limitations of the existing solutions, this paper proposes a robust and efficient approach for image spam detection using machine learning algorithm. Our proposed system analyzes the file features together with the visual features of the embedded image. These features are used to train a classifier based on back propagation neural networks to detect the email as spam or legitimate one. Experimental evaluation demonstrates the effectiveness of the proposed system comparable to the existing models for image spam classification.