2017 IEEE International Symposium on Multimedia (ISM) (2017)
Dec. 11, 2017 to Dec. 13, 2017
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/ISM.2017.112
As of February 2014 there were around 10 million registered user for massive open online courses(MOOCS) in about 1500 courses. One of the many reasons for such a huge user base is the advancements in the media streaming and storage technologies. Multimedia has emerged as a powerful and expressive medium for presenting information. However, according to a study, the average completion of MOOCS is around 15%. Though across the globe there are various options being considered to improve the completion percentage, one novel approach presented in this paper is in making the multimedia files viewer friendly to decrease consumption time. Here, we present an interactive media player that enables the user to perform audio content based searching capabilities within a multimedia file. The media player built accepts text based keyword and searches for the utterance of the entered keyword in the audio stream of the multimedia file. Markers on all successive searches are provided to ensure ease of navigation from one instance to another within the video. Our approach makes use of the keyword spotting technique available in PocketSphinx speech recogniser, which performs quick search and retrieval.
A. Chaudhary, A. K, K. Kodlekere and S. J, "Keyword Based Indexing of a Multimedia File," 2017 IEEE International Symposium on Multimedia (ISM), Taichung, Taiwan, 2017, pp. 573-576.