Key result
A multidomain speech feature extraction scheme combined with a model ensemble outperformed the state of the art by 12.3% and 33.3% in predicting heart rate and respiration, respectively.
Why the study?
Stress levels assess human well-being, and speech signals may be used to infer related physiological markers such as heart rate and respiration cycles.
Does a multidomain speech feature extraction and ensemble model improve the prediction of heart rate and respiration from speech compared to state-of-the-art methods?
Observational (n=69)
Does a multidomain speech feature extraction and ensemble model improve the prediction of heart rate and respiration from speech compared to state-of-the-art methods?
A multidomain speech feature extraction scheme combined with an ensemble model effectively predicts heart rate and respiration from speech, outperforming existing state-of-the-art methods.
May support remote cardiorespiratory monitoring from speech; leaves open prospective clinical validation.
Stress levels are a significant source of information in assessing human well-being, including both mental and physical health. Interestingly, speech signals can be indicative of stress and may be used to infer related physiological markers, such as heart rate and respiration cycles. To this end, this article proposes a nonintrusive, low-cost, and automatic stress monitoring framework facilitating timely activation of stress relief methods and/or stress prevention. Initially, we design a multidomain speech feature extraction scheme able to reveal complementary stress-related characteristics. Subsequently, these are modeled by a synergistic framework able to encode both linear and nonlinear relationships via suitably learned support vectors and a recurrent neural network. We employ an appropriate corpus encompassing recordings of job interviews, constructed based on a standardized experimental protocol. Importantly, such an approach outperformed the state of the art by 12.3% and 33.3% in predicting heart rate and respiration, respectively.
No takes yet. Share an insight, caveat, or question.
Stavros Ntalampiras (2023) conducted an observational in Stress (n=69). Multidomain feature extraction and model ensemble (SVR + ESN) vs. EGEMAPS + LSTM was evaluated on Root Mean Square Error (RMSE) for heart rate prediction on the test set (12.3% improvement). A multidomain speech feature extraction scheme combined with a model ensemble outperformed the state of the art by 12.3% and 33.3% in predicting heart rate and respiration, respectively.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: