Ontology-driven Annotation and Access of Educational Video Data in E-learning

In the past decade, we have witnessed unprecedented advances in multimedia technology. As a result, an unprecedented amount of multimedia data is being generated. Among the myriad types of multimedia data, presentation videos from lectures, conferences and seminars, and corporate trainings are of particular interest to this chapter. The need for specific solutions in this field comes from the popularity of e-learning systems. Recent years, there have been extensive efforts at both universities and colleges on developing e-learning systems to support distant learning. Based on Sloan Consortium survey (Allen & Seaman, 2008), over 3.9 million students were taking at least one online course during the fall 2007, which accounts for over 20% of all U.S. higher education students. Online enrolments continue to grow at rates far in excess of the total higher education student population, with the most recent data demonstrating no signs of slowing down. For example, the online enrolment growth rate for 2007 is 12.9%, while the growth rate for the overall higher education student population is only 1.2% for the same period (Allen & Seaman, 2008). In addition, there are e-learning systems for military, medical, and cooperate trainings (Smith, Ruocco, & Jansen, 1999; Fan, Luo, & Elmagarmid, 2004). For example, Microsoft supported 367 on-line training lectures with more than 9000 online viewers in the year of 1999 alone (He, Grudin, & Gupta, 2000). These e-learning systems enhance learning experiences and augment teachers' work in and out of traditional classrooms (Abowd, Brotherton, & Bhalodai, 1998; Flachsbart, Franklin, & Hammond, 2000). Working professionals as well embrace e-learning programs due to their convenience and flexibility (Kariya, 2003). However, due to unstructured and linear features of videos, the essential instructional content of most e-learning systems, the presentation videos, has not been fully exploited. People often feel difficulties in locating a specific piece of information in a presentation video. Sometimes they have to play back and forth several times to locate the right spot. To ensure effective exploitation of these video assets, efficient and flexible access mechanisms must be provided. Video annotation data play a critical role in video systems. The richer the annotation data are, the more flexible the video access becomes, and thus the more effective the video data can be utilized. We view video annotation as a two-step process: video segmentation, and 16