At present, three general techniques are used for noisyspeech recognition: model compensation, speech enhancement and high robustcharacteristics extraction.
目前,国际上噪声环境下的语音识别主要有3种方法:语音增强方法、模型补偿法、提取鲁棒性好的语音特征参数方法。
参考来源 - 基于小波分析方法的含噪声语音信号增强技术研究This "model compensation" is really equivalet to the "feedback linearization" for the nonlinear system.
这种”模型补偿”实质上等价于非线性系统的“反馈线性化”。
参考来源 - 非线性状态误差反馈控制律──NLSEF·2,447,543篇论文数据，部分数据来源于NoteExpress
The paper studies the appropriate objects that ADRC applies in, and extends the technology of ADRC to get model compensation ADRC and ADRC with the form in series.
研究自抗扰控制器适用的对象，并将自抗扰控制技术扩展，得到模型补偿自抗扰控制器和串联型自抗扰控制器。
Based on feature processing and model compensation, a weighted features compensation transformation method based on GMM for robust speaker verification was proposed.
从特征值处理和模型补偿两方面考虑，提出了基于高斯混合模型的加权特征补偿变换的抗噪声方法。
The collaboration engine provides a compensation model.
协作引擎提供了补偿模型。
应用推荐