ISCA Archive SMM 2023
ISCA Archive SMM 2023

Voice Technology to Identify Fatigue from Japanese Speech

Raymond Brueckner, Misa Takegami, Namhee Kwon, Nate Blaylock, Vinod Subramanian, Eri Kiyoshige, Soshiro Ogata, Yuriko Nakaoku, Henry O’Connell, Kunihiro Nishimura

Toward an automatic health monitoring tool, we investigate voice analysis technology to extract related features from speech and to build a machine learning model for identifying fatigue in Japanese. We collect voice data and their fatigue labels through phone calls and then experiment with diverse machine learning methods using various acoustic and prosodic features. The models are trained on spontaneous Japanese speech from participants who are older than 70 years. Each model and feature shows different performance and the logistic regression model using x-vectors trained on English outperforms other models with sensitivity at 0.87 and specificity at 0.65.

doi: 10.21437/SMM.2023-7

Cite as: Brueckner, R., Takegami, M., Kwon, N., Blaylock, N., Subramanian, V., Kiyoshige, E., Ogata, S., Nakaoku, Y., O’Connell, H., Nishimura, K. (2023) Voice Technology to Identify Fatigue from Japanese Speech. Proc. SMM23, Workshop on Speech, Music and Mind 2023, 31-35, doi: 10.21437/SMM.2023-7

  author={Raymond Brueckner and Misa Takegami and Namhee Kwon and Nate Blaylock and Vinod Subramanian and Eri Kiyoshige and Soshiro Ogata and Yuriko Nakaoku and Henry O’Connell and Kunihiro Nishimura},
  title={{Voice Technology to Identify Fatigue from Japanese Speech}},
  booktitle={Proc. SMM23, Workshop on Speech, Music and Mind 2023},