Transcription method and device for court hearing audio and video

Abstract

The invention relates to a court hearing audio and video transcription method and device, and the method comprises the steps: carrying out the slicing of an audio and video in a court hearing process, and generating an initial text of slices according to the sequence of the slices; performing specification conformity analysis on the initial text of the slices to obtain a conformity score of each predetermined specification; selecting a plurality of error correction tools from the error correction tool set according to the conformity scores corresponding to the predetermined specifications, and taking the selected error correction tools as target error correction tools; performing planning combination on the selected target error correction tools to obtain an error correction tool combination; performing error correction on the initial text by utilizing an error correction tool combination to obtain a sliced error correction text; and generating a court trial record text according to the error correction text of each slice. Through intelligent error correction tool selection and combination, the accuracy and normalization of court trial audio and video transcription are remarkably improved, and the method can be widely applied to automatic generation of court trial records.

Images