Critical Articulators Identification from RT-MRI of the Vocal Tract

Several technologies, such as electromagnetic midsagittal articulography (EMA) or real-time magnetic resonance (RTMRI), enable studying the static and dynamic aspects of speech production. The resulting knowledge can, in turn, inform the improvement of speech production models, e.g., for articulatory speech synthesis, by enabling the identification of which articulators and gestures are involved in producing specific sounds. The amount of data available from these technologies, and the need for a systematic quantitative assessment, advise tackling these matters through data-driven approaches, preferably unsupervised, since annotated data is scarce. In this context, a method for statistical identification of critical articulators has been proposed, in the literature, and successfully applied to EMA data. However, the many differences regarding the data available from other technologies, such as RT-MRI, and language-specific aspects create a challenging setting for its direct and wider applicability. In this article, we address the steps needed to extend the applicability of the proposed statistical analyses, initially applied to EMA, to an existing RT-MRI corpus and test it for a different language, European Portuguese. The obtained results, for three speakers, and considering 33 phonemes, provide phonologically meaningful critical articulator outcomes and show evidence of the applicability of the method to RT-MRI.