公开(公告)号：US09953634B1

公开(公告)日：2018-04-24

申请号：US14573846

申请日：2014-12-17

申请人： Knowles Electronics, LLC

发明人： David Pearce ,  Brian Clark

IPC分类号： G10L15/06

CPC分类号： G10L15/063

摘要： Provided are methods and systems for passive training for automatic speech recognition. An example method includes utilizing a first, speaker-independent model to detect a spoken keyword or a key phrase in spoken utterances. While utilizing the first model, a second model is passively trained to detect the spoken keyword or the key phrase in the spoken utterances using at least partially the spoken utterances. The second, speaker dependent model may utilize deep neural network (DNN) or convolutional neural network (CNN) techniques. In response to completion of the training, a switch is made from utilizing the first model to utilizing the second model to detect the spoken keyword or the key phrase in spoken utterances. While utilizing the second model, parameters associated therewith are updated using the spoken utterances in response to detecting the keyword or the key phrase in the spoken utterances. User authentication functionality may be provided.

公开(公告)号：US09437188B1

公开(公告)日：2016-09-06

申请号：US14667650

申请日：2015-03-24

申请人： Knowles Electronics, LLC

发明人： Eitan Asher Medina ,  David Pearce

IPC分类号： G10L15/00 ,  G10L21/00 ,  G10L15/08

CPC分类号： G10L15/08 ,  G10L15/22 ,  G10L21/0208

摘要： Systems and methods for assisting automatic speech recognition (ASR) are provided. An example system includes a buffer operable to store sensor data. The sensor data includes an acoustic signal, the acoustic signal representing at least one captured sound. The system includes a processor communicatively coupled to the buffer and being operable to store received sensor data in the buffer. The received sensor data is analyzed to produce new parameters associated with the sensor data. The buffered sensor data is processed based at least on the new parameters. The processing may include separating clean voice from noise in the acoustic signal. The processor is further operable to provide at least the processed sensor data (for example, the clean voice) to an ASR system operable to receive and process the processed sensor data at a speed faster than real time. The new parameters may also be provided to the ASR system.

摘要翻译： 提供了用于辅助自动语音识别（ASR）的系统和方法。 示例系统包括可操作以存储传感器数据的缓冲器。 传感器数据包括声信号，声信号表示至少一个捕获的声音。 系统包括通信地耦合到缓冲器并且可操作地将接收的传感器数据存储在缓冲器中的处理器。 分析所接收的传感器数据以产生与传感器数据相关联的新参数。 至少基于新参数来处理缓冲的传感器数据。 该处理可以包括将干净的声音与声信号中的噪声分开。 处理器还可操作以至少将经处理的传感器数据（例如，干净的语音）提供给可操作以以比实时更快的速度接收和处理经处理的传感器数据的ASR系统。 新参数也可以提供给ASR系统。