A Novel Approach to Speech Signal Segmentation Based on Empirical Mode Decomposition to Assess Human Psycho-Emotional State

A novel technology for speech signal segmentation into voiced, unvoiced and pause sections has been developed. It is based on the adaptive decomposition of short-time fragments of speech signals with a subsequent analysis of the parameters of empirical modes. A segmentation method based on a novel technology with a brief mathematical description is presented. The research has been conducted using the formed base of speech signals, recorded from 220 male and female subjects with signs of psycho-emotional disorders. In accordance with the obtained results, it was concluded that the method based on the developed technology determines the boundaries of informative sections more accurately, and it can be successfully tested at the preprocessing stages in the systems for detection and assessment of human psycho-emotional disorder.