G10L 15/20

Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech