Book cover Multimedia Video-Based Surveillance Systems pp 71–83Cite as Dynamic Shape Detection for Multiple Camera Systems Lucio Marcenaro & Carlo S. Regazzoni Chapter 165 Accesses 2 Citations Part of the The Springer International Series in Engineering and Computer Science book series (SECS,volume 573) Abstract Several research efforts during the past few years have focused on nonrigid object tracking [1,2]. This research field is interesting because it has several relevant applications: for example, a system able to detect movements of non-rigid objects is useful in tracking and understanding human motion. Among the main tools used to find and track a non-rigid object there are “active shape models” [3], i.e. dynamic models able to describe certain object shapes in their spatial and temporal evolutions. By using a 2D model, a 3D object shape projected onto the image plane can be characterized in order to follow appropriately the temporal evolution of the object, the model has to reflect its nature. On the basis of the nature of the objects to be modeled, active shape models can be divided into: • snakes (characterization of rigid or non-rigid object shapes) [4]; • deformable templates (based on a mathematical models of objects) [5]; • dynamic contours (based on dynamical mathematical models of objects) [6].
Dynamic Shape Detection for Multiple Camera Systems
Marcenaro, Lucio;Regazzoni, Carlo S.
2000-01-01
Abstract
Book cover Multimedia Video-Based Surveillance Systems pp 71–83Cite as Dynamic Shape Detection for Multiple Camera Systems Lucio Marcenaro & Carlo S. Regazzoni Chapter 165 Accesses 2 Citations Part of the The Springer International Series in Engineering and Computer Science book series (SECS,volume 573) Abstract Several research efforts during the past few years have focused on nonrigid object tracking [1,2]. This research field is interesting because it has several relevant applications: for example, a system able to detect movements of non-rigid objects is useful in tracking and understanding human motion. Among the main tools used to find and track a non-rigid object there are “active shape models” [3], i.e. dynamic models able to describe certain object shapes in their spatial and temporal evolutions. By using a 2D model, a 3D object shape projected onto the image plane can be characterized in order to follow appropriately the temporal evolution of the object, the model has to reflect its nature. On the basis of the nature of the objects to be modeled, active shape models can be divided into: • snakes (characterization of rigid or non-rigid object shapes) [4]; • deformable templates (based on a mathematical models of objects) [5]; • dynamic contours (based on dynamical mathematical models of objects) [6].
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
