Automatic Subtitle File-Based Sectionalization of Documentary Content and Automatic Generation of Subtitles

Abstract

The invention discloses a subtitle-file-based documentary content automatic segmentation and subhead automatic generation method and belongs to the field of Chinese text information processing. The method is composed of two phases of content automatic segmentation and subhead automatic generation and includes, in the content automatic segmentation phase, according to the time interval between neighboring subtitles recorded in the subtitle file of a documentary, selecting a plurality of positions with the largest intervals as the segmentation points to segment the entire documentary into a plurality of chapters in the time series; in the chapter subhead automatic generation phase, according to the chapter starting and stopping time determined in the previous phase, automatically selecting out a phrase combination with a grammatical structure of 'attributes and central word' from the narration of the corresponding chapter as the subhead of the chapter under the guidance of certain selecting rules. The subtitle-file-based documentary content automatic segmentation and subhead automatic generation method relatively obviously shortens the time of segmenting a relatively long documentary according to semantics through a manual manner and greatly improves the enthusiasm of teachers for taking well-made documentary segments as teaching materials.

Description

technical field [0001] The invention belongs to the field of Chinese text processing, and in particular relates to an automatic subtitle file-based subtitle file-based automatic sectioning of documentary content and an automatic generation method for subtitles. Background technique [0002] Modern, high-level classroom teaching activities require teachers to extensively collect and carefully organize various teaching materials according to teaching objectives. National Geographic, Discovery Channel, China Central Television's Science and Education Channel, Documentary Channel and other well-known documentary brand columns are undoubtedly one of the important sources of high-quality teaching materials. The use of such multimedia teaching materials can not only enrich the teaching content of teachers, but also increase students' classroom impression and stimulate students' interest in learning. [0003] However, the length of documentaries is generally more than 45 minutes, a...

AI Technical Summary

Problems solved by technology

The more complicated thing is to segment according to the change of the scene or the switching of the lens

However, the existing methods are limited by their requirements (for example, video segmentation is to facilitate the replacement of the video server cache or to extract a specific piece of content such as football shooting), and there are still two prominent defects: one is not conducive to semantic expression , that is, unless the user has watched it from the beginning to the end, the user does not know what kind of content the segmented video has; the second is not conducive to content search, that is, unless the user has watched it from the beginning to the end, the user does not know whether the segmented video has the content Other content you care about

Images