next up previous contents index
Next: Interactive dialogue systems Up: Assessing speaker verification and Previous: Sampling rare events in

Employing expert judgments to augment speaker verification and assessment for forensic aspects of speaker verification and recognition

One important application of speaker verification  is in forensic science. The forensic data that might be used can be obtained in a number of ways: Experts might listen to the material and offer a subjective judgment. This judgment is basically whether the speaker is or is not a particular speaker (i.e. discrete measures based on nominal or ordinal scales). There are lots of ways that expert judgments about speakers could be standardised which would enhance the judgments in the ways that, for example, doctors knowledge in formulating diagnosis has been done and incorporated into automatic disease diagnosis expert systems. It is beyond the scope of the current chapter to present this information. Another approach that has been taken to speaker recognition and verification has been through spectrographic measures (a continuous measure, obtained from an interval scale of measurement). To use all these measurements together calls for statistical techniques which can deal with mixes of continuous (parametric) and discrete (non-parametric) measures.

Generalised LInear Modelling Techniques (GLIM)  allow models to be constructed that involve predicting dependent variables   like those required here from mixture of discrete (including binary), and continuous measures [Aitkin et al. (1989)]. These have not been applied to forensic applications of speaker verification and recognition but seem appropriate for the task. The technique as it would be applied here enables the experimenter to establish which acoustic and subjective factors differentiate one speaker from another. In contrast, Analysis of Variance (ANOVA)  can only deal with continuous measures.    

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