Towards Biometric Person Identification using fNIRS


We investigate the potential of using fNIRS signals for biometric person identification. Independent sessions for training and testing have been recorded using 8 channels of frontal fNIRS. We extract logarithmic power spectral densities as features to train and test a Nave Bayes Classifier. We evaluate different frequency bands and report classification results for different trial lengths.

Bci Meeting 2013