Myo Keylogging Dataset

This record contains the Myo Keylogging Dataset from the paper "My(o) Armband Leaks Passwords: An EMG and IMU Based Keylogging Side-Channel Attack" by Matthias Gazzari, Annemarie Mattmann, Max Maass and Matthias Hollick in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, Volume 5, Issue 4, 2021. The dataset contains sensor data from two Myo armbands recorded from volunteers typing predefined tasks on a physical keyboard (one worn on the left and the other worn on the right forearm). Additionally, the keystrokes and their timings are included along with metadata describing each individual recording. The dataset is split into two parts, each consisting of a series of recording sessions in the respective folder: Part One: train-data In the first part, the participants were instructed to type during a series of 24 tasks consisting of copying text, pangrams, random characters and playing a game. In our paper, we used this dataset for training our classifiers. Part Two: test-data In the second part, the participants were instructed to copy four different types of passwords, namely insecure common passwords, passphrases inspired by xkcd, pwgen generated pseudo-random passwords and random passwords. In our paper, we used this dataset to evaluate our classifiers. Please refer to the README.md files in the respective folders and our paper for further details about the dataset and the data study.