Contents of this page are obsolete. This page is preserved and stored at this URL just from historical viewpoint. Original URL was http://www.mm.media.kyoto-u.ac.jp/members/kameda/...
Please visit www.kameda-lab.org for recent information. (2002/12/06, [email protected])

Publication, Research, Docs and Info, KAMEDA


next up previous
Next: Experiment Up: Video Skimming for Previous: Video Skimming for

Video Skimming

Different from video on TV, video taken by a static camera which places in a scene is boring for most of time because events which people are interested in seldom happen in front of the camera.

So the first step to handle such kind of scene watching is to remove the subsequences within which there are no events. We have developed a video skimming program that makes long video sequence shorter. It estimates area of pixels which are different from those of previous frames, and only when the area is larger than a threshold value, the frame is stored and is used to generate a digital video.

As our method assumes there is usually no movement, we cannot apply this method to a crowded corridor or pathway in downtown where a lot of people go by. Continuous movement of flow can be eliminated from the image plane by introducing analysis of optical flow, so we think we can improve our method for that situation in our future research.



Yoshinari Kameda
Tue Oct 5 01:12:48 JST 1999