Video Summary

Streaming and Adaptive Video Summary Based on Object Characterization and Spatio-Temporal Segmentation Tools

Kelvin Moutet, Marc Décombas

The number of video surveillance cameras has increased a lot during the last years. Now, it allows us to keep under surveillance public and private areas like transportation hubs, jails, city's streets, etc. Nevertheless, actually, we need to process the huge amount of videos to help human operators to keep working with them. Thanks to video processing & analysis like people detection, action recognition or video summary, all those videos can be used easier and faster.

This work provides a new approach to get a video summary based on a new concept of temporal reduction with tools especially designed or selected for streaming video summary.The proposed method will have the advantage to keep the temporal continuity of the video and increase the rate of temporal reduction while preserving important objects from the original video. Moreover, it allows us to make a streaming video summary. All of it will greatly improve the utility of the video summary tools particularly in video surveillance

________________________________________________________________________________________

Example available with ViSOR video

 R. Vezzani, R. Cucchiara, "Video Surveillance Online Repository (ViSOR): an integrated framework", Multimedia Tools and Applications, vol. 50, n. 2, Kluwer Academic Press, pp. 359-380, 2010

Visor_1261565143617_cesta4.wmv - Original Video

Video_summary_tetris_fusionST.avi - Video Summary Spatio-Temporal Reduction

Video_summary_tetris_fusionSTstreaming.avi - Video Summary Spatio-Temporal Reduction (Made in Streaming)

________________________________________________________________________________________