In this paper we propose a novel video segmentation algorithm that is capable of reliably detecting both types of dissolves in addition to cuts and wipes based on the analysis of ...
Embedding generic shape information into probabilistic spatiotemporal video object segmentation is of pivotal importance to achieving better segmentation, since it provides valuab...
Human action video sequences can be considered as nonlinear dynamic shape manifolds in the space of image frames. In this paper, we address learning and classifying human actions ...
In the paper, an automatic spatio-temporal video sequence segmentation algorithm is proposed. To address this very di cult computer vision problem, several novel algorithms have b...
In this paper, a novel shape and texture error concealment technique for segmented object-based video scenes is proposed. This technique is different from existing concealment tec...