HTTP-proxy-assisted automatic video indexing for e-learning

This paper proposes an effective approach for the construction of on-line educational video materials. We have made use of HTTP proxy through which Web-based lecture slides are accessed. The timing information stored in the proxy can be used when dividing the lecture video into shots and generating image indexes automatically. We have implemented a video recording system with segmentation and indexing module and a browsing interface to demonstrate the idea. The processing architecture and implementation overcome the problems of time-consuming manual editing and indexing.