Abstract— This paper presents a method for detecting independent temporally-persistent motion patterns in image sequences. The result is a description of the dynamic content of a...
This paper proposes a method to recover the embedding
of the possible shapes assumed by a deforming nonrigid
object by comparing triplets of frames from an orthographic
video se...
Vincent Rabaud (University of California, San Dieg...
Abstract. We introduce a novel behavioral model to describe pedestrians motions, which is able to capture sophisticated motion patterns resulting from the mixture of different cate...
A new method for object tracking in video sequences is presented. This method exploits the benefits of particle filters to tackle the multimodal distributions emerging from clutte...
Alexandros Makris, Dimitrios I. Kosmopoulos, Stavr...
3D scene understanding is key for the success of applications such as autonomous driving and robot navigation. However, existing approaches either produce a mild level of understa...