|Subject||Where Computer Vision meets Video Compression|
|Conference/Event||New AI Seminar|
|Address||University of New South Wales, Australia|
|Organisation||School of Computer Science and Engineering|
|Abstract||This is a series of seminars provided as a research overview on Multimedia Signal Processing. This short course is running over 2 sessions, 1 hour for each session on 16th and 23 Mar 2007.|
This seminar is jointly hosted by NICTA/CSE UNSW.
Digital compression of video has made a dramatic impact in the video communication domain. Existing compression standards such as MPEG-2 and MPEG-4 have been instrumental for the success of many video applications. Such applications include Digital Video Broadcast (DVB), DVD, High Definition DVD, digital camcorders and digital video on the web (i.e. YouTube).
The prime goal in the compression research community is to further squeeze the bitrate of video data for storage and transmission. In recent years more and more research concentrated on inventing and applying sophisticated Image Processing/Computer Vision algorithms to the compression domain. As such the field of video compression seems to be progressing from a primarily stochastic-based pixel approach towards a content- and region-based approach. In my talk I will give an overview about work conducted at TU Berlin in this exciting and promising field.
|Key words||Digital Video Broadcast (DVB), DVD, High Definition DVD, digital camcorders and digital video on the web|