This contribution describes an alternative approach to generating stereoscopic content of sports scenes from regular broadcast cameras without the need for special stereo rigs. This is achieved by using 3D reconstruction previously developed for applications in postmatch analysis. The reconstruction method requires at least four or five cameras and computes the 3D information automatically. Two different target formats for the delivery of stereoscopic 3D TV are discussed: the display-independent layered depth video format and conventional binocular stereo. The results viewed on two different displays demonstrate the potential of the method as an alternative production method for stereoscopic 3D content.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below