Abstract

The initialization process of a 3D model-based trackingmethod with an extremely fast checking time is presented.One initialized, the algorithm will execute a 3D model-basedtracking with pose determination. The tracking method is atwo-steps approach, a prediction-correction method. We arestudying the possibility of finding the minimum group of interestpoints and a set of rules which form all the possiblemovements of a 3D tracked object. Thereby, the computationtime-memory for the tracking task could be reduced. We considerrigid objects which can be approximately modeled by aconvex polyhedral shape. Our approach works in monocularvideos and where the position of the camera is fixed.

Reference

Torres Garcia, F., Kropatsch, W., & Artner, N. (2012). Initializing 3D Model-based Tracking of Dice. In 2012 19th International Conference on Systems, Signals and Image Processing (IWSSIP) (pp. 338–341). http://hdl.handle.net/20.500.12708/54281