No abstract

No abstract

No abstract

This paper presents the MUST-VIS system for the MediaMixer/VideoLectures.NET Temporal Segmentation and Annotation Grand Challenge. The system allows users to visualize a lecture as a series of segments represented by keyword clouds, with relations to other similar lectures and segments. Segmentation is performed using a multi-factor algorithm which takes advantage of the audio (through automatic speech recognition and word-based segmentation) and video (through the detection of actions such as writing on the blackboard). The similarity across segments and lectures is computed using a content-based recommendation algorithm. Overall, the graph-based representation of segment similarity appears to be a promising and cost-effective approach to navigating lecture databases.

Abstract-Online education is becoming more and more prevalent these days. Many universities provide pre-recorded classroom lectures for distance learning and remote users can access these lectures over Internet. With the available indexing techniques, users can search and retrieve videos related to their topic of interest in these stored databases. However, sometimes the 'mode of teaching' impacts the viewer's perception for the retrieved video lecture or snippet. In this work we make use of visual concepts in the video lecture to identify the mode of teaching and generate annotations for the video. The developed approach uses low-level features like color and edges to classify video frames into high level semantic concepts. The system performs frame-by-frame classification and mode of teaching can be inferred for each segment as well as the complete video. Experimental results show high accuracy of proposed method and demonstrate its potential for relevant applications.