A Spoof Detecting Fingerprint Reader Based on a Combination of Total Internal Reflection and Direct Image Capture

This study aims to develop a Spoof Resistant Fingerprint Reader based on a combination of Total Internal Reflection and Direct Image Capture. The system is built using a Raspberry Pi and uses two cameras connected to a multi-camera adapter to capture the picture of the fingerprint. To thoroughly examine the minutiae of the fingerprint, the project uses a series of methods using computer vision to enhance the Histogram Equalization in digital image processing. The data for this involves 60 fingerprint molds constructed by using 3 different material combinations. These being Hot Glue-Glue, Candle Wax-Wood Glue, and Clay-Gelatin. The Hot glue & Glue and Candle Wax & Wood Glue molds were tested 5 times per mold while the Clay & Gelatin mold was tested 3 times because it melts quickly. The system garnered a 95% accuracy in determining whether a fingerprint was a spoof or not.