This repository contains code to reproduce the paper:
M. Nerini, E. Favarelli, M. Chiani, "Machine learning for PIN side-channel attacks based on smartphone motion sensors," IEEE Access, 2023.
The dataset.txt file contains a 5400 x 17 matrix in which:
- each row is a sampled digit.
- each column is a feature: the first is the pressed digit and the following are motion sensor values.
The PIN_recognition.ipynb Jupiter Notebook contains the code to replicate the results in the paper.
The files rf-prod.csv, svm-prod.csv, mlp-prod.csv, and knn-sum.csv have been obtained throguh PIN_recognition.ipynb and are attached for an easier replication of the figures.