Recognition model construction method, voice recognition method, electronic equipment and storage medium

• Main Authors: JIAN RENXIAN, XU YAOQI, LIN CHANGZHOU
• Format: Patent
• Language: Chinese; English
• Published: 06.04.2021
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The invention discloses a recognition model construction method, a voice recognition method, electronic equipment and a storage medium; the method comprises the following steps: obtaining an accent extractor which is used for extracting accent features in voice; obtaining mandarin corpora of different accent; extracting accent features of the mandarin corpus by utilizing an accent extractor; and inputting the accent features and the frequency domain features of the mandarin corpus into an initial recognition model, and training the initial recognition model to obtain a recognition model . According to the invention, the accent characteristics of the voice are added into the voice recognition, so that the recognition rate of the voice with the accent can be very high. 本发明公开了识别模型构建方法、语音识别方法、电子设备及存储介质，所述识别模型构建方法包括以下步骤：获取口音提取器，所述口音提取器用于提取语音中的口音特征；获取不同口音的普通话语料；利用口音提取器提取所述普通话语料的口音特征；将普通话语料的口音特征和频域特征输入初始识别模型，对该该初始识别模型进行训练，得到识别模型。包括。本发明通过将语音的口音特征加入到语音识别中，对带口音的语音能够具有很高的识别率。