A potentially challenging technology – it is evident that a big portion of our communication is based on voice in our everyday lives.
An easy to accept method, which requires full cooperation of the user. A major disadvantage of this technology is that no word is said twice in the exact same fashion, not to mention that the human voice is easily distorted by illness, or emotional changes.
On top of that, the samples are easy to manipulate. During the identification process, the user has to say the word(s) out loud, where eavesdropping is a serious safety hazard since the machine accepts recorded and played back materials.
Wall mounted entry control devices was a failed attempt, the main current research direction is identification using phone lines. For example, a bank’s customer is identified not only by his PIN, but also by voice recognition.
While there are A number of voice recognition products introduced, most of these have faced serious issues in practice whether it is a problem with the converter or with external noise. Also, enrollment processes are more complicated than with other biometric systems. These circumstances led to voice recognition being rather unpopular, nevertheless, there is still significant work being done in this area, and its development deserves further attention.