Skip to Main Content
Presents a new human verification approach using photoplethysmography (PPG) signals that can be obtained easily from the fingertip. The experiment was performed on a group of 17 healthy subjects, and evident characteristics of the PPG signals were studied. These characteristics are unique identifiers specific to different persons while they are similar enough to recognize the same person. Four feature parameters were extracted from digitalized PPG signals. The template feature vector was formulated using part of the recorded signals, and then the discriminant function was applied to the remaining data for verification. The result with a successful rate of 94% shows that this method is promising for human verification.
Date of Conference: 24-26 April 2003