Multimedia courseware retrieval system based on voice keyword recognition

Abstract

The invention provides a multimedia courseware retrieval system based on voice keyword recognition. Firstly, a backstage converts pre-provided text knowledge points into voice models, courseware is labeled by a voice recognizing technique based on a hidden markov model to locate the accurate positions of the knowledge points in the multimedia courseware, and a reverse index based on keywords is constructed and maintained in an index module. When a user inputs text keywords in a prompt box to inquire, results in the index are extracted to be displayed to the user if the keywords are previously labeled. If the keywords are not labeled, the system can retrieve the courseware in real time, waits for feedback of the user to the results and makes statistics for the feedback information. Self-adaption training is carried out on the keywords to label the courseware again and upgrade the indexes. Compared with a traditional network course learning style system, the courseware retrieval system can quickly search for and locate the keywords of the knowledge points, improves retrieval accuracy through user interaction and finally improves the learning efficiency of students effectively.

Description

technical field [0001] The invention relates to the field of multimedia technology, in particular to a keyword-based audio retrieval method and system for multimedia courseware. Background technique [0002] In recent years, with the advancement of technology, a large number of voices have been collected and recorded and preserved in the form of audio. Among them, due to the needs of lifelong learning and network teaching, more and more multimedia audio courseware has appeared on the Internet. However, with the rapid increase of multimedia courseware, learners are faced with a large amount of teaching audio materials. If they want to find the knowledge points they want to learn, they often need to watch the materials they want to find what they need. This creates a lot of trouble for learners. Therefore, how to help learners quickly and accurately retrieve multimedia courseware and find the knowledge points they need has become an urgent problem to be solved, which involves...

AI Technical Summary

Problems solved by technology

[0004] In view of the fact that the existing technology cannot quickly and accurately find the content of the user's query in the multimedia courseware, and cannot meet the purpose of the user's efficient learning of remote courses, the technical problem to be solved by the present invention is to provide a multimedia courseware audio retrieval based on keywords Methods: Using speech recognition technology based on Hidden Markov Model to achieve accurate word standards in the set, using dynamic template regularization algorithm to achieve fast retrieval of out-of-set words, and combining user feedback to improve the accuracy of multimedia courseware retrieval

Images