Stress detection with Slovak dataset using different speech features. In this paper, we sought to build a deep learning model that would accurately classify the stress levels of Slovak speakers from acted speech corpus.
Four Jupyter Notebook files are provided. Each file presents a step-by-step procedure for extracting MFCC, PLP, TRILL, and X-vector features and training the model.
J. Kejriwal, Š. Beňuš and M. Trnka, "Stress detection using non-semantic speech representation," 2022 32nd International Conference Radioelektronika (RADIOELEKTRONIKA), Kosice, Slovakia, 2022, pp. 1-5, doi: 10.1109/RADIOELEKTRONIKA54537.2022.9764916.