Hearable technologies such as Bluetooth earbuds provides unique opportunity for stress detection as well as intervention strategies based on music listening for stress management and regulation. However, photoplethysmography (PPG) signals recorded from ear canals are often very noisy due to head movement and fit issues. In this work, we propose a heart rate variability (HRV) features estimation pipeline for PPG signals recorded using prototype Samsung Earbuds. We used template matching to determine the signal quality for dynamic fine filtering around estimated heart rate. We also improve the inter-beat interval (IBI) outlier detection and removal algorithm using global-local context of the input PPG signal. We were able to reduce the error in HRV features significantly using the proposed pipeline.

https://ieeexplore.ieee.org/document/10447778