Image prediction for virtual environments by means of Kalman filter based 3D object tracking

In this paper a new method to increase the frame rate of computer generated image sequences by means of image prediction is discussed. In general, a virtual environment consists of 3D objects, whose next position and orientation is predicted by the algorithm presented here. The objects are projected and mapped for which the texture information of the previous frame is taken into account. The usage of the previous texture information reduces a lot of computational effort and in addition latency can be reduced considerably.