Automatic scene generation using voice and text offers a unique multimedia approach to classic storytelling and human computer interaction with 3D graphics. In this paper, we pre...
The motion field of a scene can be used for object segmentation and to provide features for classification tasks like action recognition. Scene flow is the full 3D motion fiel...
Local appearance models in the neighborhood of salient image features, together with local and/or global geometric constraints, serve as the basis for several recent and effective...
Abstract. This paper presents a new approach to the problem of simultaneous location and segmentation of object in images. The main emphasis is done on the information provided by ...
A scene containing multiple independently moving, possibly occluding, rigid objects is considered under the weak perspective camera model. We obtain a set of feature points tracke...