Automatic Annotation of Instructional Videos

Abstract

Multimedia has a significant role in communicating the information and a large amount of multimedia repositories make the browsing, retrieval and delivery of video contents. For higher education, using video as a tool for learning and teaching through multimedia application is a considerable promise. To extract the audio information from the visual content is a very challenging because it needs the extraction of high level semantic information from low level visual data. The summarization technique for videos has been proposed to improve the browsing faster for large video collections to produce more efficient content indexing and access. The proposed video summarization system for instructional videos initially separates videos into audio track and then converted into text transcript. The text transcripts are pre-processed and extracted the important relevant keywords which can be used for indexing the video file. The automatic annotation of instructional videos reduced the summary length and also preserved the semantic of the videos. The research indicates that the students found these summarized content is very helpful for preparing and reviewing the lecture material and also for the preparation during their examination.