next up previous contents index
Next: Examples Up: A taxonomy of speaker Previous: Interaction mode with the

Definitions

 

Applicant speaker:  The speaker using a speaker recognition system  at a given instant. Alternative terms: current speaker, test speaker, unknown speaker,gif customer, user, etc.

Registered speaker:  A speaker who belongs to the list of registered users for a given speaker recognition system .gif For speaker classification  systems, we propose the term conform speaker (conformant speaker) to qualify a speaker who belongs to one of the classes of speakers for a given speaker classification  system.gif Alternative terms: reference speaker, valid speaker, authorised speaker, subscriber, client, etc.

Genuine speaker:  A speaker whose real identity is in accordance with the claimed identity. By extension: a speaker whose actual character and claimed class are in accordance.gif Alternative terms: authentic speaker, true speaker, correct speaker, etc.

Impostor (speaker) : In the context of speaker identification , an impostor is an applicant speaker  who does not belong to the set of registered speakers . In the context of speaker verification , an impostor is a speaker whose real identity is different from his claimed identity. Alternative terms: impersonator, usurpat ,gif etc. For speaker classification  tasks, this concept is better rendered by the term: discordant speaker.gif  

Identity assignment : Decision outcome which consists in attributing an identity to an applicant speaker , in the context of speaker identification. For speaker classification , the term class assignment should be used instead.

Acceptance : Decision outcome which consists in responding positively to a speaker (or speaker class) verification task.

Rejection : Decision outcome which consists in refusing to assign a registered identity (or class) in the context of open-set speaker identification   (or classification), or which consists in responding negatively to a speaker (class) verification trial.

(Speaker) misclassification:   Erroneous identity assignment to a registered speaker  in speaker identification .

False (speaker) rejection:gif  Erroneous rejection of a registered speaker  or of a genuine speaker  in open-set speaker identification   or speaker verification  respectively.

False (speaker) acceptance:gif  Erroneous acceptance of an impostor in open-set speaker identification   or in speaker verification .

Mistaken speaker:   The registered speaker   owning the identity assigned erroneously to another registered speaker  by a speaker identification  system.

Violated speaker:   The registered speaker owning the identity assigned erroneously to an impostor in open-set speaker identification   system. The registered speaker owning the identity claimed by a successful impostor, in speaker verification.

Text-dependent speaker recognition system:   A speaker recognition system for which the training  and test  speech utterances are composed of exactly the same linguistic material, in the same order (typically, a password).

Text-independent speaker recognition system:   A speaker recognition system for which the linguistic content of test speech utterances  varies across trials.

Personal-password speaker recognition system:   A text-dependent speaker recognition system  for which each registered speaker  has his own voice password.

Common-password speaker recognition system:   A text-dependent speaker recognition system for which all registered speakers have the same voice password.

Fixed-vocabulary  speaker recognition system:  A text-independent speaker recognition system for which test utterances  are composed of words, the order of which varies across speakers and sessions, but for which all the words were pronounced at least once by the speaker when he registered to the system.

Event-dependent speaker recognition system:   A text-independent speaker recognition system  for which test utterances  must contain a certain linguistic event (or class of events) while the rest of the acoustic material is discarded. This approach requires a preliminary step for spotting and localising the relevant events.

Unrestricted text-independent speaker recognition system:   A text-independent speaker recognition system  for which no constraints apply to the linguistic content of the test speech material .

Text-prompted  speaker recognition system:   A speaker recognition system for which, during the test phase , a written text is prompted (through an appropriate device) to the user, who has to read  it aloud.

Voice-prompted  speaker recognition system:   a speaker recognition system for which, during the test phase , the user has to repeat a speech utterance, which he listens to through an audio device.

Unprompted  speaker recognition system:   A speaker recognition system using totally spontaneous speech , i.e. for which the user is totally free to utter what he wants,gif or for which the system has no control over the speaker.gif



next up previous contents index
Next: Examples Up: A taxonomy of speaker Previous: Interaction mode with the

EAGLES SWLG SoftEdition, May 1997. Get the book...