We use Bayesian maximum a posteriori estimation training a speaker model from background model, to solve the problem of model miss matching in speaker verification system.
采用贝叶斯最大后验概率估计的方式，从统一背景模型中生成说话人模型。
Miss distance parameters are estimated by matching pursuit algorithm to search the basis function which best matches the echo signal.
通过匹配追踪算法寻找与回波信号最匹配的基函数来确定脱靶量参数。
In addition, we set up a PBMT Engine based on this algorithm. And the Engine involves the pattern-selected to avoid the problems mat come from the miss-matching of patterns.
同时我们构建了基于文本-模板直接匹配的翻译引擎，并在引擎中引入模板选优模块解决模板误匹配问题。
应用推荐