Stress level detection using double-layer subband filter

Stress level detection is important for human error prevention and health care services. Speech based stress level detection is the most effective as speech data can be obtained in nonintrusive and inexpensive ways. In this paper, we explore the features that use Double-Layered Subband (DLS) filter for detecting stress level from speech. Spectral Centroid Frequency (SCF) and Spectral Centroid Amplitude (SCA) are acoustic features that complement each other if these two are fused together using appropriate weighting coefficient. We extract SCA using DLS filters. And, we present how DLS filter integrates SCF information to SCA feature without actually computing SCF feature parameters. We investigate the effectiveness of proposed approach over combining SCA and SCF using weighting coefficient. We build user independent stress level detection system. Stress is detected using the scale of 0 to 1 (’0’ being index for no stress and ’1’ being index the highest stress level). The experiments show that the proposed system is able to detect the level of stress from speech with reasonably high accuracy.