Resumen de Applying 3D Techniques to Compute Views Provided by Video-Cameras

Roberto Yus Peirote, Eduardo Mena Nieto , Sergio Ilarri

• There exist contexts where managing several video-cameras is required, for example in the broadcasting of sports events. To identify the objects that a certain camera views (e.g., when there are multiple objects in the focus) or obtaining the kind of view (front, top, etc.) that a camera is capturing of an object are challenging tasks if we want to perform them automatically.In this paper, we describe a set of 3D techniques that can be applied to obtain the view of a camera. More specifically, they can be used to obtain the objects viewed by a camera, the kind of view of an object retrieved by the camera, and the percentage of the object or the view of an object that the camera is capturing. These techniques use a 3D model of the scenario and a 3D engine. Many dynamic factors are taken into account, such as the location of interesting objects in the environment, camera parameters, and possible occlusions, among others.