基于连续隐马尔柯夫模型的模式识别技术及其应用

对利用基因算法训练连续隐马尔柯夫模型的语音识别的具体算法进行系统的研究;然后基于该语音识别技术对深圳市司法局社区矫正声纹识别系统进行详细设计。该系统上线后的运行结果表明,利用基因算法训练连续隐马尔柯夫模型的语音识别算法的识别速度较快同时具有较高的识别率。基于模式识别技术的司法社区矫正声纹识别系统建设在我国司法系统目前尚处于起步阶段,推广和建设司法社区矫正声纹识别系统具有重要的现实意义。

Pattern Recognition Technology Based on Continuous Hidden Markov Model and its Application

Systematic research was done on the specific algorithm for speech recognition in using genetic algorithm to train continuous hidden Markov mode. Then the detailed design of Voiceprint Recognition System of Community Correction Objects in the Shenzhen City Bureau of Justice has been done based on the speech recognition technology. The system running results show that the recognition rate of recognition algorithm using genetic algorithm to train continuous hidden Markov model is faster and has a higher rate of recognition. Construction of voiceprint recognition system of judicial community correction objects based on pattern recognition is still in the junior stage in our judicial system, and promotion and the construction of voiceprint recognition system of judicial community correction objects have the important practical significance.