Using dysphonic voice to characterize speaker's biometry

Phonation distortion leaves relevant marks in a speaker's biometric profile. Dysphonic voice production may be used for biometrical speaker characterization. In the present paper phonation features derived from the glottal source (GS) parameterization, after vocal tract inversion, is proposed for dysphonic voice characterization in Speaker Verification tasks. The glottal source derived parameters are matched in a forensic evaluation framework defining a distance-based metric specification. The phonation segments used in the study are derived from fillers, long vowels, and other phonation segments produced in spontaneous telephone conversations. Phonated segments from a telephonic database of 100 male Spanish native speakers are combined in a 10-fold cross-validation task to produce the set of quality measurements outlined in the paper. Shimmer, mucosal wave correlate, vocal fold cover biomechanical parameter unbalance and a subset of the GS cepstral profile produce accuracy rates as high as 99.57 for a wide threshold interval (62.08-75.04%). An Equal Error Rate of 0.64 % can be granted. The proposed metric framework is shown to behave more fairly than classical likelihood ratios in supporting the hypothesis of the defense vs that of the prosecution, thus ofering a more reliable evaluation scoring. Possible applications are Speaker Verification and Dysphonic Voice Grading.

Development of a genre-dependent TTS system with cross-speaker speaking-style transplantation

One of the biggest challenges in speech synthesis is the production of contextually-appropriate naturally sounding synthetic voices. This means that a Text-To-Speech system must be able to analyze a text beyond the sentence limits in order to select, or even modulate, the speaking style according to a broader context. Our current architecture is based on a two-step approach: text genre identification and speaking style synthesis according to the detected discourse genre. For the final implementation, a set of four genres and their corresponding speaking styles were considered: broadcast news, live sport commentaries, interviews and political speeches. In the final TTS evaluation, the four speaking styles were transplanted to the neutral voices of other speakers not included in the training database. When the transplanted styles were compared to the neutral voices, transplantation was significantly preferred and the similarity to the target speaker was as high as 78%.

New methodology for the integration of biometric features in speaker recognition systems applied to security environments

La cuestión principal abordada en esta tesis doctoral es la mejora de los sistemas biométricos de reconocimiento de personas a partir de la voz, proponiendo el uso de una nueva parametrización, que hemos denominado parametrización biométrica extendida dependiente de género (GDEBP en sus siglas en inglés). No se propone una ruptura completa respecto a los parámetros clásicos sino una nueva forma de utilizarlos y complementarlos. En concreto, proponemos el uso de parámetros diferentes dependiendo del género del locutor, ya que como es bien sabido, la voz masculina y femenina presentan características diferentes que deberán modelarse, por tanto, de diferente manera. Además complementamos los parámetros clásicos utilizados (MFFC extraídos de la señal de voz), con un nuevo conjunto de parámetros extraídos a partir de la deconstrucción de la señal de voz en sus componentes de fuente glótica (más relacionada con el proceso y órganos de fonación y por tanto con características físicas del locutor) y de tracto vocal (más relacionada con la articulación acústica y por tanto con el mensaje emitido). Para verificar la validez de esta propuesta se plantean diversos escenarios, utilizando diferentes bases de datos, para validar que la GDEBP permite generar una descripción más precisa de los locutores que los parámetros MFCC clásicos independientes del género. En concreto se plantean diferentes escenarios de identificación sobre texto restringido y texto independiente utilizando las bases de datos de HESPERIA y ALBAYZIN. El trabajo también se completa con la participación en dos competiciones internacionales de reconocimiento de locutor, NIST SRE (2010 y 2012) y MOBIO 2013. En el primer caso debido a la naturaleza de las bases de datos utilizadas se obtuvieron resultados cercanos al estado del arte, mientras que en el segundo de los casos el sistema presentado obtuvo la mejor tasa de reconocimiento para locutores femeninos. A pesar de que el objetivo principal de esta tesis no es el estudio de sistemas de clasificación, sí ha sido necesario analizar el rendimiento de diferentes sistemas de clasificación, para ver el rendimiento de la parametrización propuesta. En concreto, se ha abordado el uso de sistemas de reconocimiento basados en el paradigma GMM-UBM, supervectores e i-vectors. Los resultados que se presentan confirman que la utilización de características que permitan describir los locutores de manera más precisa es en cierto modo más importante que la elección del sistema de clasificación utilizado por el sistema. En este sentido la parametrización propuesta supone un paso adelante en la mejora de los sistemas de reconocimiento biométrico de personas por la voz, ya que incluso con sistemas de clasificación relativamente simples se consiguen tasas de reconocimiento realmente competitivas. ABSTRACT The main question addressed in this thesis is the improvement of automatic speaker recognition systems, by the introduction of a new front-end module that we have called Gender Dependent Extended Biometric Parameterisation (GDEBP). This front-end do not constitute a complete break with respect to classical parameterisation techniques used in speaker recognition but a new way to obtain these parameters while introducing some complementary ones. Specifically, we propose a gender-dependent parameterisation, since as it is well known male and female voices have different characteristic, and therefore the use of different parameters to model these distinguishing characteristics should provide a better characterisation of speakers. Additionally, we propose the introduction of a new set of biometric parameters extracted from the components which result from the deconstruction of the voice into its glottal source estimate (close related to the phonation process and the involved organs, and therefore the physical characteristics of the speaker) and vocal tract estimate (close related to acoustic articulation and therefore to the spoken message). These biometric parameters constitute a complement to the classical MFCC extracted from the power spectral density of speech as a whole. In order to check the validity of this proposal we establish different practical scenarios, using different databases, so we can conclude that a GDEBP generates a more accurate description of speakers than classical approaches based on gender-independent MFCC. Specifically, we propose scenarios based on text-constrain and text-independent test using HESPERIA and ALBAYZIN databases. This work is also completed with the participation in two international speaker recognition evaluations: NIST SRE (2010 and 2012) and MOBIO 2013, with diverse results. In the first case, due to the nature of the NIST databases, we obtain results closed to state-of-the-art although confirming our hypothesis, whereas in the MOBIO SRE we obtain the best simple system performance for female speakers. Although the study of classification systems is beyond the scope of this thesis, we found it necessary to analise the performance of different classification systems, in order to verify the effect of them on the propose parameterisation. In particular, we have addressed the use of speaker recognition systems based on the GMM-UBM paradigm, supervectors and i-vectors. The presented results confirm that the selection of a set of parameters that allows for a more accurate description of the speakers is as important as the selection of the classification method used by the biometric system. In this sense, the proposed parameterisation constitutes a step forward in improving speaker recognition systems, since even when using relatively simple classification systems, really competitive recognition rates are achieved.

Glottal Parameter Estimation by Wavelet Transform for Voice Biometry

Voice biometry is classically based on the parameterization and patterning of speech features mainly. The present approach is based on the characterization of phonation features instead (glottal features). The intention is to reduce intra-speaker variability due to the `text'. Through the study of larynx biomechanics it may be seen that the glottal correlates constitute a family of 2-nd order gaussian wavelets. The methodology relies in the extraction of glottal correlates (the glottal source) which are parameterized using wavelet techniques. Classification and pattern matching was carried out using Gaussian Mixture Models. Data of speakers from a balanced database and NIST SRE HASR2 were used in verification experiments. Preliminary results are given and discussed.

Towards an unsupervised speaking style voice building framework: multi-style speaker diarization

Current text-to-speech systems are developed using studio-recorded speech in a neutral style or based on acted emotions. However, the proliferation of media sharing sites would allow developing a new generation of speech-based systems which could cope with spontaneous and styled speech. This paper proposes an architecture to deal with realistic recordings and carries out some experiments on unsupervised speaker diarization. In order to maximize the speaker purity of the clusters while keeping a high speaker coverage, the paper evaluates the F-measure of a diarization module, achieving high scores (>85%) especially when the clusters are longer than 30 seconds, even for the more spontaneous and expressive styles (such as talk shows or sports).

Consulta on-line basada en corpus de construcciones con preposición en el texto técnico en inglés y su traducción al español

Basándonos en la recopilación inicial de preposiciones, locuciones preposicionales, términos con preposición dependiente y phrasal verbs utilizados en el texto técnico realizada en otros proyectos anteriores del Departamento de Lingüística Aplicada a la Ciencia y a la Tecnología, el objetivo de este trabajo es completar, organizar, actualizar y dar visibilidad a esta información inicial. Tras realizar un proceso exhaustivo de verificación, unificación, clasificación y ampliación de la información existente, en caso necesario, el listado resultante se utiliza para elaborar un glosario de términos con preposición. El objetivo final de este proyecto es que este glosario esté a disposición de los usuarios, a través de una consulta on-line, en la página del ILLLab (http://illlab.euitt.upm.es/wordpress/), dependiente del Departamento de Lingüística Aplicada a la Ciencia y a la Tecnología. Para incluir en el glosario ejemplos actualizados de textos técnicos, se ha recopilado un corpus lingüístico de textos técnicos, tomando como base diferentes números de la revista IEEE Spectrum, en su edición digital, publicados entre los años 2009 y 2012. El objetivo de esta recopilación es la de ofrecer al consultante diferentes ejemplos de uso en el texto técnico de los distintos términos con preposición que componen el glosario, de manera que pueda acceder de manera rápida y sencilla a ejemplos de uso real de los términos que está buscando, con objeto de clarificar aspectos relacionados con su uso o, en su caso, facilitar su aprendizaje. Toda esta información, tanto el listado de términos con preposición como las frases pertenecientes al corpus recopilado, se incorpora a una base de datos, alojada dentro de la misma página web del ILLLab. A través de un formulario de consulta, a disposición del usuario en dicha página, se pueden obtener todos los términos recopilados que coincidan con los criterios de búsqueda introducidos. El usuario puede realizar dos tipos de búsqueda principales: por preposición o por término completo. Además, puede elegir una búsqueda global (entre todos los términos que integran el glosario) o parcial (en una sola de las categorías en las que se han dividido los diferentes términos, de acuerdo con su función gramatical). Por último, se presentan unas estadísticas de uso de los términos recopilados dentro de los diferentes textos que integran el corpus lingüístico, de manera que pueda establecerse una relación de los que aparecen con más frecuencia en el texto técnico. ABSTRACT. Based on the initial collection of prepositions, prepositional phrases, dependent prepositions and phrasal verbs used in technical texts collected on previous projects in the Department of Applied Linguistics to Science and Technology, the aim of this project is to improve, organize, update and provide visibility to this initial information. Following a process of verification, unification, classification and extension of existing information, if necessary, a glossary of terms with preposition is built. The ultimate objective of this project is to make this glossary available to users through an online consultation in the ILLLab webpage (http://illlab.euitt.upm.es/wordpress/). The administration of tis webpage depends of the Department of Applied Linguistics in Science and Technology. A linguistic corpus of technical texts has been compiled, based on different numbers of the IEEE Spectrum magazine, in its online edition, published between the years 2009 and 2012. The aim of this collection is to provide different examples of use in the technical text for the terms included in the glossary, so that examples of the actual use of the terms consulted can be easily and quickly accessed, in order to clarify doubts regarding their meaning or translation into Spanish and facilitate learning. All this information, both the list of terms with prepositional phrases as well as the corpus developed, is incorporated in a database. Through a searching form, the ILLLab's user may obtain all the terms matching the search criteria entered. The user can perform two types of main search: by preposition or by full term. Additionally, a global search can be selected (including all terms included in the glossary) or a partial one (including only one of the glossary's categories). Finally, some statistics of use are presented according to the various texts included in the corpus, so a relation of the most frequent prepositions in the technical text can be established.

Region-dependent vehicle classification using PCA features

Video-based vehicle detection is the focus of increasing interest due to its potential towards collision avoidance. In particular, vehicle verification is especially challenging due to the enormous variability of vehicles in size, color, pose, etc. In this paper, a new approach based on supervised learning using Principal Component Analysis (PCA) is proposed that addresses the main limitations of existing methods. Namely, in contrast to classical approaches which train a single classifier regardless of the relative position of the candidate (thus ignoring valuable pose information), a region-dependent analysis is performed by considering four different areas. In addition, a study on the evolution of the classification performance according to the dimensionality of the principal subspace is carried out using PCA features within a SVM-based classification scheme. Indeed, the experiments performed on a publicly available database prove that PCA dimensionality requirements are region-dependent. Hence, in this work, the optimal configuration is adapted to each of them, rendering very good vehicle verification results.

Classical vs. biometric features in the 2013 speaker recognition evaluation in mobile environments

MFCC coefficients extracted from the power spectral density of speech as a whole, seems to have become the de facto standard in the area of speaker recognition, as demonstrated by its use in almost all systems submitted to the 2013 Speaker Recognition Evaluation (SRE) in Mobile Environment [1], thus relegating to background this component of the recognition systems. However, in this article we will show that selecting the adequate speaker characterization system is as important as the selection of the classifier. To accomplish this we will compare the recognition rates achieved by different recognition systems that relies on the same classifier (GMM-UBM) but connected with different feature extraction systems (based on both classical and biometric parameters). As a result we will show that a gender dependent biometric parameterization with a simple recognition system based on GMM- UBM paradigm provides very competitive or even better recognition rates when compared to more complex classification systems based on classical features

Histograms of oriented gradients for fast on-board vehicle verification

Histograms of Oriented Gradients (HoGs) provide excellent results in object detection and verification. However, their demanding processing requirements bound their applicability in some critical real-time scenarios, such as for video-based on-board vehicle detection systems. In this work, an efficient HOG configuration for pose-based on-board vehicle verification is proposed, which alleviates both the processing requirements and required feature vector length without reducing classification performance. The impact on classification of some critical configuration and processing parameters is in depth analyzed to propose a baseline efficient descriptor. Based on the analysis of its cells contribution to classification, new view-dependent cell-configuration patterns are proposed, resulting in reduced descriptors which provide an excellent balance between performance and computational requirements, rendering higher verification rates than other works in the literature.