Analysis System Of Teaching Behavior In Education Videos

Since the concept “MOOC” appeared, people from all over the world benefits from the qualified online education resources and enjoys the equal chances to accept education. To some extent, it relaxes the temporal and spatial limit of traditional education, and appeases the problems of unbalanced distribution. However, the open course videos vary a lot in quality, and the appropriate or suitable ones are hardly found by learners. To address this problem, some methodologies of video processing and analyzing could be utilized to study the educational behaviors inspired by the assessment technologies of traditional teaching in classroom. On the other side, massive open course videos provide data for the research of education. It will greatly contribute to promoting teaching behaviors and educational quality. Considering that, we attempt to analyze the teaching behaviors and find the reasonable quantitative behavioral description in open course videos via the technologies of video processing and analyzing instead of the traditional methods based on references and questionnaire statistics.We firstly revisit the progress of the research of teaching behaviors and briefly introduce and analyze some state-of-the-art associated methods and frameworks. Then, based on Flanders Interaction Analysis System(FIAS), we recognize the identities of speakers, subtitle, projecting curtain, and classify the teaching behaviors in three aspects, i.e., the speech and the usage of multimedia assistance. Finally, we design and implement an online learning platform. It could support retrieving, browsing, organizing and combining educational materials, meanwhile, provide some tools for manual annotation of teaching behaviors, speech and identities of speakers.In summary, we attempt to employ the technologies of video processing and analyzing to objectively and quantitatively describe the teaching behaviors in open course videos. The preliminary experiment results show that our solution is reasonable and yet has the room to make promotion. The online learning platform and some annotation tools developed on WEB 2.0 works well and stably.