You cannot create a working system based on hashing an initial sample, then comparing hashed values of a live sample. Due to the nature of fingerprints and the readers, you cannot count on getting the same exact image (or corresponding minutae measurements) when sampling the same finger.

Therefore, you cannot compare hashes because they would only match if the input is exactly the same.

Modern fingerprint algorythms compare the stored fingerprint (template) to the live sample and “score” the result. If the result meets a specified threshold, then it is called a match.

So Privaris, like others, stores an encrypted file called a template containing measurements of the fingerprint key characteristics (minutia).

Works great, and as you correctly said these minutia based templates cannot be reverse engineered to create a fingerprint image.

Jeff