Image Classifiers for Network Intrusions
This addresses network security by detecting intrusions, but is incremental as it applies existing methods to a new dataset representation.
This research transforms network intrusion data into grayscale images and applies MobileNetV2 to achieve 97% accuracy in binary classification of normal vs. attack traffic, with 56% accuracy for 9 specific attack families.
This research recasts the network attack dataset from UNSW-NB15 as an intrusion detection problem in image space. Using one-hot-encodings, the resulting grayscale thumbnails provide a quarter-million examples for deep learning algorithms. Applying the MobileNetV2's convolutional neural network architecture, the work demonstrates a 97% accuracy in distinguishing normal and attack traffic. Further class refinements to 9 individual attack families (exploits, worms, shellcodes) show an overall 56% accuracy. Using feature importance rank, a random forest solution on subsets show the most important source-destination factors and the least important ones as mainly obscure protocols. The dataset is available on Kaggle.