A comparison of procedures for the calculation of forensic likelihood ratios from acoustic-phonetic data: Multivariate kernel density (MVKD) versus Gaussian mixture model-universal background model (GMM-UBM)

Two procedures for the calculation of forensic likelihood ratios were tested on the same set of acoustic-phonetic data. One procedure was a multivariate kernel density procedure (MVKD) which is common in acoustic-phonetic forensic voice comparison, and the other was a Gaussian mixture model-universal background model (GMM-UBM) which is common in automatic forensic voice comparison. The data were coefficient values from discrete cosine transforms fitted to second-formant trajectories of /a@?/, /e@?/, /o@?/, /a@?/, and /@?@?/ tokens produced by 27 male speakers of Australian English. Scores were calculated separately for each phoneme and then fused using logistic regression. The performance of the fused GMM-UBM system was much better than that of the fused MVKD system, both in terms of accuracy (as measured using the log-likelihood-ratio cost, C"l"l"r) and precision (as measured using an empirical estimate of the 95% credible interval for the likelihood ratios from the different-speaker comparisons).

[1]  A. Agresti.  An introduction to categorical data analysis , 1997 .

[2]  Julien Epps,et al.  Estimating the Precision of the Likelihood-Ratio Output of a Forensic-Voice-Comparison System , 2010, Odyssey.

[3]  Franco Taroni,et al.  Statistics and the Evaluation of Evidence for Forensic Scientists , 2004 .

[4]  Sadaoki Furui,et al.  International Speech Communication Association , 2006 .

[5]  J. Buckleton,et al.  Forensic DNA Evidence Interpretation , 2004 .

[6]  Andrzej Drygajlo,et al.  Scoring and direct methods for the interpretation of evidence in forensic speaker recognition , 2004, INTERSPEECH.

[7]  Philip Rose,et al.  Realistic Extrinsic Forensic Speaker discrimination with the Diphthong / ai/ , 2006 .

[8]  Didier Meuwly,et al.  Forensic speaker recognition based on a Bayesian framework and Gaussian mixture modelling (GMM) , 2001, Odyssey.

[9]  J M Curran,et al.  Assessing uncertainty in DNA evidence caused by sampling effects. , 2002, Science & justice : journal of the Forensic Science Society.

[10]  David Lucy,et al.  Introduction to Statistics for Forensic Scientists , 2005 .

[11]  Eliathamby Ambikairajah,et al.  FM features for automatic forensic speaker recognition , 2008, INTERSPEECH.

[12]  David A. van Leeuwen,et al.  Fusion of Heterogeneous Speaker Recognition Systems in the STBU Submission for the NIST Speaker Recognition Evaluation 2006 , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[13]  Phil Rose,et al.  Technical forensic speaker recognition: Evaluation, types and testing of evidence , 2006, Comput. Speech Lang..

[14]  Philip Rose,et al.  Linguistic-Acoustic Forensic Speaker Identification with Likelihood Ratios from a Multivariate Hierarchical Random Effects Model - A Non-Idiot's Bayes' Approach , 2004 .

[15]  Geoffrey Stewart Morrison,et al.  Automatic-type calibration of traditionally derived likelihood ratios: forensic analysis of australian English /o/ formant trajectories , 2008, INTERSPEECH.

[16]  Law. Policy.  Executive Summary of the National Academies of Science Reports, Strengthening Forensic Science in the United States: A Path Forward , 2009 .

[17]  Ian W. Evett,et al.  Statistical analysis of STR data , 1996 .

[18]  I. Evett,et al.  A hierarchy of propositions: deciding which level to address in casework , 1998 .

[19]  D. Hosmer,et al.  Applied Logistic Regression , 1991 .

[20]  Elizabeth Shriberg,et al.  An Anticorrelation Kernel for Subsystem Training in Multiple Classifier Systems , 2009, J. Mach. Learn. Res..

[21]  Douglas A. Reynolds,et al.  Speaker Verification Using Adapted Gaussian Mixture Models , 2000, Digit. Signal Process..

[22]  Richard S. Frank,et al.  The Theory of Interpreting Scientific Transfer Evidence , 1990 .

[23]  J A Lambert,et al.  The impact of the principles of evidence interpretation on the structure and content of statements. , 2000, Science & justice : journal of the Forensic Science Society.

[24]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[25]  J A Lambert,et al.  A model for case assessment and interpretation. , 1998, Science & justice : journal of the Forensic Science Society.

[26]  Colin Aitken,et al.  Corrigendum: Evaluation of trace evidence in the form of multivariate data , 2004 .

[27]  G. Morrison.  Likelihood-ratio forensic voice comparison using parametric representations of the formant trajectories of diphthongs. , 2009, The Journal of the Acoustical Society of America.

[28]  Geoffrey Stewart Morrison,et al.  Forensic voice comparison and the paradigm shift. , 2009, Science & justice : journal of the Forensic Science Society.

[29]  J. Curran.  An introduction to Bayesian credible intervals for sampling error in DNA profiles , 2005 .

[30]  Pascal Druyts,et al.  Applying Logistic Regression to the Fusion of the NIST'99 1-Speaker Submissions , 2000, Digit. Signal Process..

[31]  Niko Brümmer,et al.  Application-independent evaluation of speaker detection , 2006, Comput. Speech Lang..

[32]  Solange Rossato,et al.  Intra-speaker variability effects on Speaker Verification performance , 2010, Odyssey.

[33]  Christian Müller,et al.  Speaker Classification I: Fundamentals, Features, and Methods , 2007, Speaker Classification.

[34]  John Buckleton.  A Framework for Interpreting Evidence , 2004 .

[35]  D. Balding.  Weight-of-Evidence for Forensic DNA Profiles , 2005 .

[36]  David Lindley,et al.  A problem in forensic science , 1977 .

[37]  Javier Ortega-Garcia,et al.  Bayesian analysis of fingerprint, face and signature evidences with automatic biometric systems. , 2005, Forensic science international.

[38]  Shunichi Ishihara,et al.  How many do we need? exploration of the population size effect on the performance of forensic speaker classification , 2008, INTERSPEECH.

[39]  Tony Alderman.  The Use of Australian-English Vowel Formant Data Sets in Forensic Speaker Identification , 2004 .

[40]  Javier Ortega-Garcia,et al.  Robust estimation, interpretation and assessment of likelihood ratios in forensic speaker recognition , 2006, Comput. Speech Lang..

[41]  Doroteo Torre Toledano,et al.  Emulating DNA: Rigorous Quantification of Evidential Weight in Transparent and Testable Forensic Speaker Recognition , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[42]  I. W. Evett,et al.  Towards a uniform framework for reporting opinions in forensic science casework , 1998 .

[43]  David G. Stork,et al.  Pattern Classification , 1973 .

[44]  Andrzej Drygajlo,et al.  Aural and automatic forensic speaker recognition in mismatched conditions , 2005 .

[45]  Daniel Ramos Castro.  Extended Abstract for Best Ph.D. Thesis Award: Forensic Evaluation of the Evidence Using Automatic Speaker Recognition Systems , 2007 .

[46]  Philip Rose,et al.  The Intrinsic Forensic Discriminatory Power of Diphthongs , 2006 .

[47]  Colin Aitken,et al.  The use of statistics in forensic science , 1991 .

[48]  David A. van Leeuwen,et al.  An Introduction to Application-Independent Evaluation of Speaker Recognition Systems , 2007, Speaker Classification.

[49]  Javier Ortega-Garcia,et al.  Likelihood Ratio Calibration in a Transparent and Testable Forensic Speaker Recognition Framework , 2006, 2006 IEEE Odyssey - The Speaker and Language Recognition Workshop.

[50]  Hirotaka Nakasone,et al.  Forensic automatic speaker recognition , 2001, Odyssey.

[51]  Catherine I. Watson,et al.  Impact of the GSM mobile phone network on the speech signal: some preliminary findings , 2009 .

[52]  Philip Rose.  Forensic Speaker Identification , 2002 .

[53]  Didier Meuwly.  Reconnaissance de locuteurs en sciences forensiques: l'apport d'une approche automatique , 2000 .

[54]  Phil Rose,et al.  Beyond the long-term mean: exploring the potential of F0 distribution parameters in traditional forensic speaker recognition , 2008, Odyssey.

[55]  Geoffrey Stewart Morrison.  Forensic voice Comparison using likelihood ratios based on Polynomial curves fitted to the formant trajectories of Australian English , 2009 .

[56]  Colin Aitken,et al.  Evaluation of trace evidence in the form of multivariate data , 2004 .

[57]  Phil Rose,et al.  A response to the UK position statement on forensic speaker comparison , 2009 .

[58]  Michael Jessen,et al.  SPES: The BKA Forensic Automatic Voice Comparison System , 2010, Odyssey.

[59]  Michael Jessen,et al.  Forensic speaker verification using formant features and Gaussian mixture models , 2008, INTERSPEECH.

[60]  Anil Alexander,et al.  Forensic automatic speaker recognition using Bayesian interpretation and statistical compensation for mismatched conditions , 2007 .

[61]  Philip Rose,et al.  Exploring the Discriminatory Potential of F0 Distribution Parameters in Traditional Forensic Speaker Recognition , 2009 .

[62]  A. Drygajlo,et al.  Forensic Automatic Speaker Recognition [Exploratory DSP] , 2007, IEEE Signal Processing Magazine.

[63]  Didier Meuwly,et al.  The inference of identity in forensic speaker recognition , 2000, Speech Commun..

[64]  Walter Bär,et al.  Advances in Forensic Haemogenetics , 1988, Advances in Forensic Haemogenetics.

[65]  Phil Rose.  Accounting for Correlation in Linguistic-Acoustic Likelihood Ratio-based Forensic Speaker Discrimination , 2006, 2006 IEEE Odyssey - The Speaker and Language Recognition Workshop.

[66]  Philip Rose,et al.  FORENSIC SPEAKER DISCRIMINATION WITH AUSTRALIAN ENGLISH VOWEL ACOUSTICS , 2007 .