UCLA EMBEDDED SECURITY GROUP
- Home - Research - People - Publications - Links - Press -


Fingerprint Signal Processing Algorithms

The raw data collected by the CMOS sensor must be processed and the biometric features must be extracted via a process known as minutiae detection. Using the standard floating-point C NIST detection algorithm as a base, we customized the algorithm for inclusion in an embedded context. Matching is a set of algorithms used to match a pre-stored template with a candidate fingerprint obtained from the sensor. In terms of security, the signal processing algorithms were designed to be embedded within the ThumbPod to avoid the sensitive biometric data being sent over the wireless channel. 

 

 

Los Angeles - UCLA