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

Levels of text dependence


Another feature which is classically used to specify a speaker recognition   system is its level of text dependence, i.e. the constraints on the linguistic material imposed on a test utterance . A main distinction is conventionally set between text-dependent systems   and text-independent systems  . Though this basic distinction is not accurate enough to cover the range of practical possibilities, below we give a definition of these two terms according to the usage found in the literature. To simplify, in text-dependent systems,   the linguistic content of the training  and test material  are totally identical, while in text-independent systems   test utterances   vary across trials (at least in terms of word order).

However, a deeper study of the various strategies used in practice shows that at least five levels of text dependence should be distinguished. Two of them resort to text-dependent approaches, but can be opposed to the use of either a personal password gif or a common password.gif   The other three can be viewed as several variants of text-independent   approaches, using either fixed words in a random order ( fixed-vocabulary systemsgif   ), a specific linguistic event, wherever it occurs (event-dependent systemsgif  ), or a completely unrestricted text (unrestricted text-independent systemsgif    ).

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