论文部分内容阅读
在吸收听觉生理学研究成果基础上,建立了一个模拟外围听觉系统和部分中枢听觉神经系统功能的听觉模型.以该模型作为前端处理来提取特征参数,后处理采用模式匹配的方法.进一步研究了听觉模型用于语音识别的抗噪声问题,根据听觉生理特点提出了几个抗噪声方案.实验结果表明,基于文中的听觉模型和所提抗噪声方案的语音识别系统,在噪声环境下具有很好的鲁棒性(Robustnes).
Based on the research results of auditory physiology, an auditory model was established to simulate the peripheral auditory system and some central auditory nervous system functions. The model is taken as the front-end processing to extract the characteristic parameters, and the post-processing is mode-matched. The anti-noise problem of auditory model for speech recognition is further studied. Several anti-noise schemes are proposed according to the characteristics of auditory physiology. Experimental results show that the speech recognition system based on the hearing model and anti-noise scheme proposed in this paper has good robustness in noisy environments.