Adaptive intelligent generation method of picture-text video thumbnails based on query terms

A thumbnail and self-adaptive technology, applied in video data retrieval, video data browsing/visualization, special data processing applications, etc., can solve problems such as multi-storage space, fuzzy key frame definition, key frame selection, etc., to save time , increase browsing efficiency and reduce storage space
CN109145152AActive Publication Date: 2019-01-04SUN YAT SEN UNIV

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
SUN YAT SEN UNIV
Publication Date
2019-01-04

Smart Images

  • Figure 1
    Figure 1
Patent Text Reader

Abstract

The invention discloses a self-adaptive intelligent generation method of picture-text video thumbnails based on query words. The method comprises the following steps: acquiring an object video, extracting the object video, and obtaining audio and video information in the video; performing structured processing on the audio and video information to obtain structured video data and structured audiodata; selecting the structured video data to obtain a key frame corresponding to the query keyword semantics, namely a visual element; extracting and processing the structured audio data to obtain text elements related to query keyword semantics; obtaining visual elements and text elements for dynamic composition processing, and obtaining picture-text video thumbnails; the thumbnails are obtained,the semantic text of the thumbnails is extracted, the global color matching monitoring process is carried out, and the target video thumbnails related to the semantics of the query keywords are obtained. The embodiment of the invention can intelligently generate video thumbnails according to query keywords according to a system, thereby saving human resources and having more purposefulness than the existing automatic generation technology of video thumbnails.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical fields of algorithm optimization and analysis, computer image processing and deep learning, in particular to a method for adaptively and intelligently generating graphic and video thumbnails based on query words. Background technique

[0002] Video thumbnails play a vital role in how users find videos and evaluate content relevance. The selection of thumbnails is not only related to the magnitude of video transmission in social networks, but also seriously affects the instant gratification of online viewers' information acquisition. However, selecting the most suitable thumbnail from thousands of video frames is a time-consuming and skillful manual process. Although there are currently some methods that can automatically or semi-automatically extract thumbnails from videos, there are still great limitations in specific scenarios, such as when the user needs to specify a video thumbnail related to a certain topic ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More