Video efficient retrieval system supporting fuzzy comment mining

Abstract

According to the invention, a high-level abstract concept of videos is extracted, a set of video retrieval system based on fuzzy comment mining is realized, comment data of various videos are collected through a coding web spider, the problem of dynamic webpage information crawling for various videos is solved, the comment data are cleaned firstly, then an association rule mining algorithm is adopted. A frequently appearing noun set is extracted from comments, a noise set is removed through isolation and correlation pruning, possible non-subject frequent words are filtered by adopting point mutual information to obtain a feature word set close to a video in order to improve the correlation between feature words and video subjects, and repeated expressions are clustered by adopting a clustering algorithm; and possible related contents of a video correspond to a series of feature nouns and related topics surrounding each feature word on the basis of an information community, topic mining is performed through an LDA topic model, and the method is realized on the basis of a Lucene open source retrieval framework and has higher accuracy and efficiency.

Description

technical field [0001] The invention relates to a high-efficiency video retrieval system, in particular to a high-efficiency video retrieval system supporting fuzzy comment mining, and belongs to the technical field of video fuzzy retrieval. Background technique [0002] With the rapid popularization of mobile smart devices, the way for people to obtain video is becoming more and more convenient. The amount of online video has also exploded. Currently, more and more people watch short videos on mobile devices. In the face of massive videos on the Internet, fast and accurate video retrieval has become a research and development focus. [0003] At present, there are two directions for video retrieval: one is keyword-based topic retrieval, which is essentially text retrieval. For video materials that need to be retrieved, they first need to be manually marked. In specific web pages, web pages The editor draws up the title of the video and adds the relevant description. The se...

AI Technical Summary

Problems solved by technology

[0013] Aiming at the deficiencies of the prior art, the present invention successfully solves the hot issue of accurately retrieving the required video in massive video data, and proposes a new video retrieval method with high efficiency, accuracy, moderate algorithm complexity and strong practicability: based on video The fuzzy retrieval of concept expansion mines and refines the high-level abstract concept information that cannot be obtained intuitively, such as the theme that needs to be expressed, the idea explained, the included roles, and the involved scenes, and expands the retrieved domain of video objects. When submitting a fuzzy abstract description, it can accurately and efficiently query, and successfully implements an efficient video retrieval system that supports fuzzy comment mining.

Images