An overview of text-independent speaker recognition: From features to supervectors

TLDR: This paper starts with the fundamentals of automatic speaker recognition, concerning feature extraction and speaker modeling and elaborate advanced computational techniques to address robustness and session variability.

About: This article is published in Speech Communication.The article was published on 2010-01-01 and is currently open access. It has received 1433 citations till now. The article focuses on the topics: Speaker recognition.

Figures

Figure 4: Glottal feature extraction [116]. Speech frame (top), linear prediction (LP) residual (middle), and glottal flow estimated via inverse filtering (bottom). c©2009 IEEE. Reprinted by permission.

Figure 9: Principle of support vector machine (SVM). A maximummargin hyperplane that separates the positive (+1) and negative (-1) training examples is found by an optimization process. SVMs have excellent generalization performance.

Figure 11: The concept of modern sequence kernel SVM. Variablelength utterances are mapped into fixed-dimensional supervectors, followed by intersession variability compensation and SVM training.

Figure 12: Example of detection error trade-off (DET) plot presenting various subsystems and a combined system using score-level fusion.

Figure 1: Components of a typical automatic speaker recognition system. In the enrollment mode, a speaker model is created with the aid of previously created background model; in recognition mode, both the hypothesized model and the background model are matched and background score is used in normalizing the raw score.

Figure 5: Extracting modulation spectrogram features [123]. A timefrequency context, including M short-term spectra over the interval [n0 . . . n0 + M − 1], is first extracted. The DFT magnitude spectra of all feature trajectories are then computed and stacked as a feature vector with high dimensionality (here 129 × 65 = 8385 elements).