Publication Date

Spring 2018

Degree Type

Master's Project


Computer Science


Image classification is a fundamental problem of computer vision and pattern recognition. Spam is unwanted bulk content and image spam is unwanted content embedded inside the images. Image spam creates threat to the email based communication systems. Nowadays, a lot of unsolicited content is circulated over the internet. While a lot of machine learning techniques are successful in detecting textual based spam, this is not the case for image spams, which can easily evade these textual-spam detection systems. In this project, we explore and evaluate four deep learning techniques that detect image spams. First, we study neural networks and the deep neural networks, which we train on various image features. We explore their robustness on an improved dataset, which was especially build in order to outsmart current image spam detection techniques. Finally, we design two convolution neural network architectures and provide experimental results for these alongside the existing VGG19 transfer learning model for detecting image spams. Our work offers a new tool for detecting image spams and is compared against recent related tools.