Abstract:

The first experiments in lecture video-recording started about 20 years ago [1]. Over the last 15 years the technique became relatively standard as a form of support of learning. Several investigations were conducted to assess it, showing that methodologies which include video-recording offer many advantages. At the same time, research has been carried on to improve the technology itself (e.g. by having the camera automatically panning to follow the lecturer): for a review see [2,3].

Most systems however lack some form of semantic indexing of the content, which turns out to be extremely useful for students, who seek a particular passage in the lecture. Our data indicate that very often students do not watch the entire lecture on-line, but rather that they search for a particular section, e.g. to check their notes or to review a critical section of a lecture.

Automatically extracting information that may be helpful to accomplish this mission is not an easy task.

We present a possible solution to this problem. Our solution is based on superimposing a marker on the slides used by the teacher to allow automatic recognition of the slides themselves. Our proposed solution is based on an unobtrusive, small QR-Code which is automatically superimposed on the (e.g. PowerPoint) slides, and on a software for detecting such marker in the recorded stream. In such way, every slide can be associated with a temporal window. The titles of the slides can then provide an index, which can be presented to the viewer to quickly find a particular spot in the lecture. Also, the content of the slides can be indexed, so as to allow for textual search: the slide spotted by the search results can then be used (via the index) to jump to the a particular point in the video, achieving a simple form of multimedia indexing.

In the paper we will present the state of the art, illustrate our idea, describe the implementation and the tests of our proposed solution and finally discuss the results.

References:
[1] Tobagi. F. “Distance learning with digital video”. Multimedia, IEEE vol. 2 (1) pp. 90 – 93 (1995)
[2] M. Ronchetti, “Video-Lectures over Internet: The Impact on Education” in E- Infrastructures and Technologies for Lifelong Learning: Next Generation Environments, IGI Global, pp. 253-270 (2011)
[3] M. Ronchetti, “Perspectives of the Application of Video Streaming to Education” in Streaming Media Architectures, Techniques, and Applications: Recent Advances, Hershey PA, USA: Information Science Reference, IGI Global, pp. 411-428 (2011)

Keywords:

Video-lectures, semantic annotation.