htn274/CanBus-IDS

Semi-supervised Deep Learning Based In-vehicle Intrusion Detection System Using Convolutional Adversarial Autoencoder

This is the implementation of the paper "Detecting In-vehicle Intrusion via Semi-supervised Learning-based Convolutional Adversarial Autoencoders"

Requirements & Libraries

• Python3
• Tensorflow 1.15

Dataset

• Car-Hacking dataset

Code

• preprocessing.py: transform csv format into tfrecord format
• train_test_split.py: split total data into train/test/validation set
• train.py: train/test the model by using is_train option
• test_performance.py: measure the inference time of the model