Article Preview
TopIntroduction
For years video has been used to record lectures for distance learning and reviewing (Brofferio, 1998). Indeed video has the power to capture and reproduce more lecture and presentation features than any other widely available media like text transcriptions or audio recordings: in addition to presenting text, illustrations and speech, video also conveys lecturer’s gestures, interactions with the audience and the classroom scene in general.
Recording and sharing lectures has also got a lot easier in recent years as there is no longer need for dedicated recording crews, invasive capture devices or specially tailored classroom environments. In sum, any lecturer can now setup a tripod-mounted camera and record his entire presentation without assistance (Soong, Chan, Cheers, & Hu, 2006). Resulting videos can then be shared using free Internet lecture and educational videos’ repositories such as iTunes U (Apple Inc., 2007) and YouTube EDU (Google Inc., 2009) among others.
Going All Digital, and Mobile
The enthusiast way our education community embraced new technologies can be first explained by the effectiveness for creating and delivering digital learning material. Additionally, with the advent of mobile platforms, reaching learners has never been this straightforward (Ormond, 2008).
A second reason why eLearning has been received so warmly is the fact that most learning material has been significantly enhanced when going digital. For example textbooks and printed material have gained in interactivity and navigability allowing students to browse text and image-based material in small chunks and in the order that best suits their learning path, all while skipping unneeded content with unprecedented ease.
Notable exceptions to such digital enhancements are recorded lecture videos. Although producing recorded lectures is more accessible and popular than ever, we cannot help wondering if at the other end students will be genuinely interested in resulting videos and their inherent media shortcomings.
Ever since becoming mainstream video has largely remained unchanged: from videotape recordings to high definition broadcasts –all conceived with entertainment content in mind– video is still a rigid and linear medium. While general entertainment videos can be watched from beginning to end sequentially, and knowledge about the actual contents and their structure is not critical, recorded lectures are not suitable for such passive viewing. After all a typical lecture repository consists of hefty video files with dozens of recorded hours of overwhelmingly indistinguishable content.
A typical video player, such as the one illustrated in Figure 1, doesn’t provide navigation cues other than an inexpressive timeline without any real useful information besides the current playing position and duration of the recording while lecture contents remain obscure. In most cases users have no other option than blindly jumping around the video, or even worse, skipping recorded lectures altogether because of frustration.
Figure 1. A recorded lecture viewed with a regular mobile video player. Notice how the very limited navigation timeline (a) presents no information about the lecture contents or structure.
With our research we intend to provide a complete solution for learners with intuitive information of recorded lecture contents to enable them to precisely find sections of interest and browse recordings without guesswork. At the same time we aim to leverage videos’ interactivity and navigation features by providing several visual cues and annotations features.
In the next section we first describe an automated summarization tool that we have developed for blackboard-based recorded lectures. Later on we present our mobile visualization tool that combines source lecture videos with our summarization analysis output to provide a novel lecture viewing experience.