Common Evaluation Pitfalls in Touch-Based Authentication Systems

Artifacts for "Common Evaluation Pitfalls in Touch-Based Authentication Systems".

Data

The data used in this paper is available on ORA.

The features.zip file contains the features.csv which is the only file needed to recreate the experiments presented in the paper.

If you would like to perform the feature extraction yourself, the data_files.zip folder contains all the necessary files with the following structure:

├── data_files                  
│   ├── UUID_X1                                    # User ID  
│   │   │── SESSION_Y1                             
│   │   │   │── scroll                             # Social media task 
│   │   │   │   │── REPETITON_W1       
│   │   │   │   │   │── accelerometer_data
│   │   │   │   │   │   └── ###-###-###.csv 
│   │   │   │   │   │── gyroscope_data
│   │   │   │   │   │   └── ###-###-###.csv 
│   │   │   │   │   └── touch_data
│   │   │   │   │       └── ###-###-###.csv 
|   |   |   |   └── REPETITON_W2    
|   |   |   |       └── ... 
│   │   │   └── swipe                               # Image gallery task
│   │   │       │── REPETITON_Z1
│   │   │       │   │── accelerometer_data
│   │   │       │   │   └── ###-###-###.csv
│   │   │       │   │── gyroscope_data
│   │   │       │   │   └── ###-###-###.csv
│   │   │       │   └── touch_data
│   │   │       │       └── ###-###-###.csv
│   │   |       └── REPETITON_Z2    
│   │   |           └── ... 
│   │   |
│   │   └── SESSION_Y2
│   │       └── ...
│   └── UUID_X2
│   │   └── ..
│   └── ... 

Using this codebase

To use this repository, you will also need docker. Then complete the following steps:

1. git clone https://github.com/ssloxford/evaluation-pitfalls-touch.git
2. cd evaluation-pitfalls-touch/data
3. wget 'https://ora.ox.ac.uk/objects/uuid:5f1abaa7-52a4-430b-9208-128d9f1832fd/download_file?safe_filename=features.zip&type_of_work=Dataset' -O 'features.zip'
   • (Optional) If you would like to generate the feature.csv yourself run:
     wget 'https://ora.ox.ac.uk/objects/uuid:5f1abaa7-52a4-430b-9208-128d9f1832fd/download_file?safe_filename=data_files.zip&type_of_work=Dataset' -O 'data_files.zip' instead.
4. unzip features.zip
   • (Optional) unzip data_files.zip
5. cd .. && docker build -t evaluation-pitfalls-touch .
6. To start and connect to the docker container use:
   docker run -it -v $(pwd)/data:/touch/data --name evaluation-pitfalls-touch evaluation-pitfalls-touch
   • (Optional) cd data && python3 feature_extraction.py && cd ..
7. You can now run the experiments - for instance cd experiments && python3 p2_single_phone_model.py

iOS Application

We also share the source code of the iOS application in the that was developed for data collection purposes. This version can be found in the ios-app folder but is not immediately functional as a back-end server is needed to make it work. However, it can be used as a reference or made usable with minor modifications.

Contributors

• Martin Georgiev
• Simon Eberz
• Henry Turner
• Giulio Lovisotto

Acknowledgements

This work was generously supported by a grant from Mastercard and by the Engineering and Physical Sciences Research Council [grant numbers EP/N509711/1, EP/P00881X/1].