基于一阶有限差分商的带噪语音端点检测方法

摘要

Endpoint detection of speech recognition technology is the basis of voice recognition,endpoint detection with low SNR environment determined largely the work of speech recognition can be conducted effectively.This paper presents a method based on first-or%语音识别中的端点检测技术是语音识别的基础,低信噪比环境下的端点检测很大程度决定语音识别工作的有效进行.提出了一种利用一阶有限差分商结合折半查找来判定语音端点检测的方法,提高了端点检测的准确率.仿真实验表明,在噪声环境下该方法较好地实现了语音端点检测.