High-performance session variability compensation in forensic automatic speaker recognition. | Publicación