Subtitle generation method and device and device for generating subtitles

Abstract

The embodiment of the invention provides a subtitle generation method and device and a device for generating subtitles. The method comprises the steps of carrying out voice recognition processing on audio and video signals to be processed, obtaining a text sequence corresponding to the audio and video signals and a timestamp mapping table of the text sequence, wherein the timestamp mapping table comprises timestamps corresponding to all segmented words in the text sequence; determining boundary segmented words in the text sequence; and splitting the text sequence according to the boundary segmented words and the timestamps corresponding to the boundary segmented words to generate subtitle line files corresponding to the audio and video signals. According to the embodiment of the invention, the accuracy and fluency of the generated subtitle file can be improved.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a method and device for generating subtitles and a device for generating subtitles. Background technique [0002] When a user is watching some audio and video, such as webcast or movie, he can understand the audio and video content through the subtitles displayed on the audio and video display screen. [0003] In traditional methods for generating audio and video subtitles, audio streams are mainly processed according to silent segments so as to generate subtitles. The silent segment may be a segment without voice in the audio stream of the audio and video. According to the silent segment, the audio stream is segmented into multiple voice segments, and then speech recognition is performed on the segmented voice segments to obtain subtitles corresponding to the voice segments. [0004] However, the method of segmenting the audio stream based on a single audio signal fea...

AI Technical Summary

Problems solved by technology

[0004] However, the method of segmenting the audio stream based on a single audio signal feature such as a silent segment is difficult to distinguish the expression pauses within a sentence and the expression pauses between sentences in a character's speech, and the semantics expressed by the segmented speech segments are often incomplete , for speech recognition of segmented speech segments, the recognition results are often not accurate enough, and the accuracy and fluency of the generated subtitles are poor, which is not conducive to users' understanding of audio and video content

Images