An Automatic Lecture Note Taker

There is growing interest in higher education for electronic learning resources and tools which could make use of the large technological infrastructure that has built up around students and institutions over the past few years. This paper investigates the technology that would be required to produce a tool for the automatic visual and aural capture of lectures. A possible implementation is then presented, based on a system of key-frames to encapsulate the lecture’s content and meaning, and is tested against synthetic, mock-, and real lecture data. The resulting system shows promising results, but is intolerant to some of the more unanticipated complexities of real video data and the evaluation concludes that while most of the initial design is effective, some is ineffective in dealing with real data.