Project: Advanced multiresolution techniques for modeling and segmentation of video
01-OCT-06 – 30-SEP-10
The proposed research aims at developing new multiresolution techniques for video segmentation and video coding. We wish to code video objects separately and in a scalable manner, where the spatial and the temporal resolution are adjusted depending on the channel and the receiver characteristics. Important nolvelites of this research will include segementation techniques based on combined textural and motion information taking into account reliability of the motion estimates, and joint coding of object geometry and texture.