Voice biometrics is an example of Polish software innovation in Life Science (IT sector in Cracow – Personal data protection)

Globe on tablet computer at office desktop. Media app icons flying around the globe. There is a warm glow behind the globe.

Innovations are a challenge to current legal solutions. For example, voice biometrics technology may cause a problem for the applicable legal provisions in personal data protection. In Poland the area of personal data protection is regulated under Personal Data Protection Act of 29 August 1997 (Journal of Laws of 2016, item 922)


What is biometrics?

Do you want to take care of something through the hotline of your bank or call centre? Or maybe you already have so much passwords, that you are not able to remember them all? Accounts, passwords, logins, other varied verification processes – taking care of safety becoming time-consuming. But there is a technology, that enables to save your time – it is called voice biometrics.

Biometrics refers to metrics related to human characteristics. Nowadays there is an increasing use of this technology, especially in the matters related to security: at the airport, where you can choose facial recognition system instead of traditional way of check in. Voice biometrics is a technology using to identify people with their voices.


Our voice is unique like our fingerprints. Moreover, all of us articulate sentences in original way: stress differently, rate of speech and intonation are different. The system records and picks out all differences with taking into consideration details such as size and shape of throat, mouth cavity, nasal cavity, length and tension of vocal cords. Every recorded voice print is stored as a mathematical model. To avoid mistakes during the voice recording, the commands are dictated by speech synthesizer. Verification process consists in comparing sample of recordings to the previous recording. What if we have a cold or sore throat? Or if the noise occurs in the background? The systems currently are equipped with technologies of removing ambient noise, therefore they can recognize our voice in most cases.

Variants of voice biometrics

There are two types of voice biometrics:

  • active voice biometrics (also called text-dependent)
  • passive voice biometrics (also called text-independent)

In the first type the text must be the same for enrolment and verification. The second type of voice biometrics does not require saying some predetermined phrase. The verification process begins when the talk with e.g. consultant starts.

Voice biometrics appears to be a course in which the security systems should head towards, since it is a precise, simple and safe method to protect your accounts as evidenced by the low rates of False Match Rate (FMR) and False Rejection Rate (FRR).