Speech and speaker recognition application on the TMS320C541 board [articol]

Abstract

The paper presents a speech and speaker recognition application developed on the EVM C541 board using the CCS®. The application represents the implementation of the TESPAR coding method on a DSP support. The TESPAR alphabet for the coding process was obtained formerly. The speech/speaker information contained in the utterances is extracted by TESPAR coder and provides the TESPAR A matrices. For the recognition decision, the distances among the TESPAR A test matrix and the TESPAR A reference matrices are computed. The results of the experiments prove the high capabilities of the TESPAR method in the classification tasks.