发送短信 : Speaker recognition based on adaptive boosting and parallel multiple expert models /