数据模型归一化，模态振动，完整的基于HMM的语音识别系统，粒子图像分割及匹配均为自行编制的子例程。- Normalized data model, modal vibration, Complete HMM-based speech recognition system, Particle image segmentation and matching subroutines themselves are prepared.