cti编程的全面例子

这是语音信号处理中求倒谱的程序,我自己编的,很简单,实现了一段序列求倒谱的功能-This is the voice signal processing for cepstrum procedures, I developed, is very simple. achieve a sequence for the functional cepstrum

隐马尔 柯夫模型( HM M ) 的传统 训练方法 ——Baum-W elch 算法只能 得到一个局 部最优 模型, 从而影 响 最终的识别率。对于 CHM M , 分段 K  平均方法来取得一个初始值可以解决这一问题, 但对 DHM M 却改进不大。 而 基于全 局搜索的 进化计 算的一 个重要 特点便 是可以 得到次 优解乃 至全 局最优 解。本文 把进 化计算 引入 到 DHM M 的训练中去, 提出一种改 进的进化训练方法, 实验结果表 明, 这种训 练方法初 具了全局搜 索和快速收 敛 的特点, 得到的模型优于传统方法和直接用进化计算所得的模型, 提高了系统的识别率。

关于SDK的语音识别和语音结合 的代码-On the SDK Speech Recognition and voice combined with the code

基于GMM的概率神经网络PNN具有良好的泛化能力，快速的学习能力，易于在线更新,并具有统计学的贝叶斯估计理论基础，已成为一种解决像说话人识别、文字识别、医疗图像识别、卫星云图识别等许多实际困难分类问题的很有效的工具。而且PNN不但具有GMM的大部分优点，还具有许多GMM没有的优点，如强鲁棒性，需要更少的训练语料，可以和其他网络其他理论无缝整合等。-GMM based probabilistic neural network PNN good generalization ability, the ability to learn fast, easy online updates, and with the Bayesian statistical theory based on estimates, and has become a solution as speaker recognition, text recognition, medical image recognition, satellite images and other real recognition when difficulties classification of very effective tool. But GMM PNN is not only the most advantages, but also has many advantages GMM not as strong robustness, require less training corpus, and other networks to other theories, such as seamless integration.

语音识别专辑四- Speech recognition special edition four

国际上最新的说话人识别算法的源程序，内置语音训练数据-The latest international speaker recognition algorithm source code, built-in voice training data

带参考信号的盲源分离，很好用的程序，用于语音增强-ICA-R

对HMM基础知识的了解很有帮助，有发展历史，应用等-The basis of HMM knowledge helpful, there are the history of the development, application, etc.

Speech Quantization_Compare SNRs of speech cs5 with mu-law companding and uniform quantization -Speech Quantization_Compare remnants of spee ch cs5 with mu-law companding and uniform quant ization

采用VC＋＋编程和G.729协议，实现语言的实时通信，内含源代码。-using VC++ programming and G.729 agreement, the realization of real-time communication language, containing the source code.