This code implements a very simple "biometric" authentication system using the
fuzzy vault algorithm as described by Juels and Sudan in
A Fuzzy Vault Scheme.
I have made some notable simplifications (in fuzzy_vault.py):
- The "biometric" data is represented as a list of ten floats. Real fingerprint data is more complex.
- I have simplified the polynomial interpolation -- rather than using Reed-Solomon codes, I used the a polynomial fit function.
- Chaff points are not thrown out if they collide with genuine points on the polynomial.
To run the program, choose a fingerprint file from the fingerprints directory,
e.g. ming and run:
python authenticate.py fingerprints/ming
This program imports a list of vaults, which were created using "fingerprints"
(biometric templates). Each vault stores encrypted fingerprint data bound with a
secret key (i.e. the name of the person). Thus, when authenticate.py is
invoked with the supplied template, we try to unlock each vault to get the
name encrypted in it. If the template is a close enough match to the original
template used to create the vault and the name returned is on a list of known
users, we get a match and the user is accepted.