Speech Recognition Based on Efficient DTW Algorithm and Its DSP Implementation

摘要

The present voice recognition has achieved high recognition accuracy in theory and the laboratory condition, but many speech recognition algorithms are realize on PC. A system platform with DSP core can realize real-time speech processing algorithms, and in cost, power consumption and volume has the advantages of PC did not own, and has a good application prospect. This paper emphasis on how to realize speech recognition in DSP and its whole hardware and software technological process, and analyzed the program with CCS's self-contained process analysis tools Profile. In order to improve the recognition rate, a high performance DTW algorithm which differ from the traditional DTW algorithm is used. So real-time control is achieved and obtain a high recognition rate.