中文摘要：

研究了G．729编码流说话人识别算法，提出了一种使用固定码书增益作为判决参数的G．729静音帧判决方法。将这一方法应用于低复杂度的G．729编码流说话人识别，去除压缩语音数据中的静音帧从而提高识别率。

英文摘要：

The algorithm of low-complexity automatic speaker recognition in the coded G.729 stream was researched and a method using the fixed codebook gain value to choose silent frames in G.729 coded speech was introduced. The method was applied to low-complexity automatic speaker recognition in the coded G.729 stream to discard the silent frames.