Hacker News Reader

Instead of storing plaintext passwords, most companies store a hash of that password, essentially applying a one-way transformation to it and saving the result of that. I don’t have to know your password, I can just apply the hash algorithm to whatever you enter in the input and if that matches, we’re golden.

No reason you couldn’t apply the same philosophy to biometric data.

View on HN

Yes, I agree on passwords since the characters are the same every time.

Does every iris scan produce the same hash though?

Didn’t say literally the same thing. Obviously it would be more complex. but I’d assume a fair amount of consistency between scans. How else would you compare at all?

Does your iris change between scans? I assume not.

This feels like a flippant response. The question you responded to was ‘would the hash of the iris would be the same?’ It isn’t as if you’ll get an identical image of the iris every time, and hashes tend to behave chaotically for even slightly different values. If we compare this to something like password salting and hashing, it isn’t clear how we can maintain the constant salted hash value if we swap the password for a digital representation of a person’s iris.

Does it need to be exactly the same? I assume it's a perceptual hash with a liveness check.

Reply on news.ycombinator.com