A conference shorthand system and conference shorthand method

Abstract

The invention discloses a conference shorthand system and a conference shorthand method. The conference shorthand system is mainly composed of a conference shorthand terminal for recording conference audio, an ASR server for providing voice recognition services, an NLP server for providing natural language processing services, and a collaborative editor providing background support. The server is composed of a manual editing terminal for correcting meeting records. The meeting shorthand terminal is connected to the ASR server, the NLP server, and the collaborative editing server in two directions, and the collaborative editing server is connected to the manual editing terminal in two directions. The conference stenography terminal cuts the audio stream according to natural sentences, which reduces the bandwidth in the audio transmission process and makes the transmission faster. The text return speed of the ASR server and the NLP server is also faster; an audio segment and its corresponding text After being transmitted to the manual editing terminal, it can be corrected according to the audio segment and its corresponding text, so as to realize the real-time correction of the dynamically generated meeting record.

Description

technical field [0001] The invention relates to the technical field of voice shorthand, in particular to a conference shorthand system and a conference shorthand method capable of real-time correction of conference records. Background technique [0002] During the meeting, the recorder records the organization and specific content of the meeting, forming the minutes of the meeting. The most traditional form is that the recorder takes on-site shorthand and checks the meeting minutes based on the meeting recordings after the meeting is over. [0003] With the development of speech recognition technology (ASR) and natural language processing technology (NLP), the audio generated in the meeting can be directly converted into text and generate meeting records in real time at the meeting site, which greatly reduces the workload of recorders. [0004] Speech recognition technology converts the lexical content in human speech into computer-readable input, such as keystrokes, binary...

AI Technical Summary

Problems solved by technology

At present, the correction method adopted is mainly that after the meeting is over, the recorder sorts out and checks the meeting minutes according to the meeting recordings, which makes the generation of the meeting minutes a certain delay and inconvenience.

It is easy to think that the optimal correction method is of course to modify the text converted from the audio in real time, but the technical obstacle is how to realize the real-time modification of the text while the audio is being recorded and the text is being generated. , Quick correction, that is, how to timely and quickly correct the text that is being dynamically generated

Images