In this paper a novel scheme for colored video compression using color transfer techniques is proposed. Color transfer for video sequences is made more accurate by incorporation of...
Local spatiotemporal features or interest points provide compact but descriptive representations for efficient video analysis and motion recognition. Current local feature extract...
In this paper we describe a method to learn parameters
which govern pedestrian motion by observing video
data. Our learning framework is based on variational
mode learning and a...
The dynamic texture (DT) is a probabilistic generative model, defined over space and time, that represents a video as the output of a linear dynamical system (LDS). The DT model ...
In this paper, temporal segmentation of 3D video based on motion analysis is presented. 3D video is a sequence of 3D models made for a real-world dynamic object. A modified shape ...