Two things needs to be done in this dataset.

• This one is a good candidate to port to the new structure vanilla-biometrics
• Needs to have data for score normalization