Toward Language-Agnostic Speech Emotion Recognition

Cross-language speech emotion recognition is receiving increased attention due to its extensive real-world applicability. This work proposes a language-agnostic speech emotion recognition algorithm focusing on Italian and German languages. We combine mel-scaled and temporal modulation spectral representations which are subsequently modeled by means of Gaussian mixture models. Emotion prediction is carried out via a Kullback Leibler divergence scheme. Importantly, we apply the proposed methodology on two problem settings, i.e. one including positive vs. negative emotion classification and a second one where all Big Six emotional states are considered. A thorough experimental campaign demonstrated the efficacy of such a method, as well as its superiority over other generative modeling schemes and state of the art approaches.