Implementation of speech recognition system based on VC++

By taking advantage of VC++, this paper has realized one kind of speaker-independent speech recognition system with small vocabulary and isolated words. By adopting the method of pre-emphasis endpoint detection, this system can eliminate effects of low-frequency noise. Moreover, it invokes HTK speech processing toolbox in VC to complete the feature extraction and pattern matching. According to the test result, this system can make real-time and accurate recognition of words in a certain vocabulary.