Biometric Voice Recognition Credit Cards 122
securitas writes "New Scientist's Celeste Biever reports on the latest in biometric security devices: voice recognition credit cards. The device is three times the size of a normal credit card, has a 'microphone, a loudspeaker, a battery and a voice-recognition chip' and is intended to help reduce credit card fraud. The owner speaks a password into the card and the card emits an authentication squawk. Bruce Schneier loves the concept of BeepCard's related sound authentication technology. Other articles at the Telegraph and The Register."
Re:3x the size!?! (Score:3, Informative)
How does it work? (Score:5, Informative)
Smart cards are similar to credit cards and serve the same purpose, but they have a completely different data storage system. Instead of using a magnetic strip to store the user's information, smart cards feature an embedded 8-bit microprocessor with up to 16 kilobytes of programmable-only memory. Smart cards have only recently began to gain popularity in the United States, despite their widespread use in Europe for years.
Domain Dynamics' new TESPAR voice authenticator stores three samples of the user's voice on a template within the Java-based smart card. When users want to make a transaction, they simply enter the card into a terminal at a store and give a speech sample. The card then matches the spoken voice to the recorded voice samples, a process that takes a fraction of the second. The company said that TESPAR is able to handle day-to-day variations in the user's voice and can ignore background noise.
Read more here [howstuffworks.com].
Re:My voice is my password (Score:3, Informative)
My voice is my passport, verify me.
Sneakers rocks.
--Bob
Re:Bubba Smith (Score:1, Informative)
-AC
Re:Bubba Smith (Score:3, Informative)
Oh what an institution!
Stupid Designer Tricks (Score:2, Informative)
No, the challenge for Beepcard has been and apparently continues to be knowing the difference between voice recognition and voice identification. Yes, the phones have to learn to recognize your voice but that's not because your voice is a beautiful and unique snowflake; it's because the parameters that make up voices are widely varying and we haven't figured out exactly how brains turn streams of voice into words. So-called "voiceprints" are a myth. Google for forensic phonetics if you don't believe me. IANAPhonetician but I am a linguistics grad student and I've had enough grad level phonetics to know that spectrographic analysis of a voice does not provide any kind of unique identification the way fingerprinting does (or is supposed to at least).
Stupid idea, waste of R&D money, and a poor solution to a problem that has many better solutions that credit card companies are not interested in because it involves them changing the way they do things. Whoever's financing this obviously needs to enroll in an introductory phonetics class at the local university before they lose their shirts on a dumb, unworkable idea.