Method and system for improving audio transfer accuracy in multi-person dialogue scene

The invention discloses a method and a system for improving voice transcription accuracy in a multi-person dialogue scene. The method comprises the following steps: acquiring to-be-recognized audio data; based on the audio data, performing small sentence segment unit division processing to obtain a small sentence segment unit division processing result; based on the small sentence segment unit division processing result, accurate recognition processing is carried out, and a recognition processing result is obtained; performing noise reduction processing based on the recognition processing result to obtain a noise reduction processing result; based on the noise reduction processing result, performing audio identification conversion processing to obtain conversion data; and front-end display is carried out based on the conversion data, and audio identification in a multi-person common dialogue audio scene is realized. According to the method and the device, the problems of long time consumption and low recognit