Principal Component Analysis Implementation for Signal Processing of Electrochemical Impedance Spectroscopy in the Detection of Fake Fingerprints

Abstract

Electrochemical impedance spectroscopy (EIS) is widely used to analyze biometric data such as medical and bio-health. This EIS method is intended to be used in the detection of human fingerprints and fake fingerprints. The most import factor in the detection of such fake fingerprints lies in the ability to increase the discrimination of fake fingerprints compared to the human fingerprints. In particular, the deviation of the EIS value varies depending on how the finger is in contact with the electrode. if the detection is not passed and retry is required. To solve this problem, principal component analysis (PCA) is applied. PCA is widely used as a method of extracting various data features. In order to effectively apply EIS signal data to PCA, the change of wave form according to 10 frequencies between 6K and 15K was generated and compared to input vs. output. After measuring the wave magnitude of the output signal and the time to reach 80% of the wave-max value, it was analyzed with PCA to determine the wave trend. And more higher discrimination was obtained than when using only the decision tree method in the detection.