摘要 |
A voice data recognition method and device for distinguishing a regional accent, and a server. The voice data recognition method for distinguishing a regional accent comprises: calculating the voice recognition confidence and/or signal-to-noise ratio of voice data, and according to the voice recognition confidence and/or signal-to-noise ratio, screening regional voice data from the voice data (S110); and according to a region attribute of the regional voice data, determining a region to which the regional voice data belongs (S120). By means of the voice data recognition method and device for distinguishing a regional accent, and the server, the regional voice data is automatically recognized from mass voice data by way of calculating the confidence and signal-to-noise ratio of the voice data, thereby avoiding manually labelling the voice data, and therefore, the processing efficiency of the voice data is improved. |