Video strip splitting method and device, electronic equipment and readable medium

Through lens detection and video content representation generation, combined with correlation, video clips are merged, which solves the problem of ignoring semantic coherence when video stripping, and achieves a more accurate video stripping effect.

CN119729149APending Publication Date: 2025-03-28BEIJING IQIYI TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411892080.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-20
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

When unpacking videos, the prior art ignores the semantic consistency and completeness of video content and cannot accurately capture the logic and coherence of video clips.

Method used

Through lens detection, the target video is divided into multiple video clips, a video content representation is generated, and the correlation degree of each group of adjacent video clips is determined based on the video content representation, and a video clip that meets the preset conditions is combined according to the correlation degree to form a segment combination.

Benefits of technology

The semantic information of the video content is fully considered, and the problem of ignoring semantic coherence and integrity when disassembling videos is solved, achieving more accurate disassembly of videos.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a video strip splitting method and device, electronic equipment and a readable medium, and the method comprises the steps: dividing a target video into a plurality of video segments through lens detection; generating a video content representation for each video segment; determining the relevancy of each group of adjacent video clips according to the video content representation; and merging the video clips meeting a first preset condition into a clip combination according to the relevancy so as to split the target video into each clip combination. The video content representation of each video clip in the target video is firstly calculated, and then whether the adjacent video clips can be combined into the same clip combination is determined according to the relevancy between the video content representations of the adjacent video clips. Therefore, the semantic information is fully considered when the video is split, and the problem that the semantic coherence and integrity of the video content are ignored when the video is split is solved.
Need to check novelty before this filing date? Find Prior Art