| | |
Summary: Model-Based Motion Estimation for Synthetic Animations
Maneesh Agrawala Andrew C. Beers Navin Chaddhay
Computer Science Department yComputer Systems Laboratory
Stanford University Stanford University
ABSTRACT
One approach to performing motion estimation on syn-
thetic animations is to treat them as video sequences
and use standard image-based motion estimation meth-
ods. Alternatively, we can take advantage of informa-
tion used in rendering the animation to guide the motion
estimation algorithm. This information includes the 3D
movements of the objects in the scene and the projec-
tion transformations from 3D world space into screen
space. In this paper we examine how to use this high
level object motion information to perform fast, accu-
rate block-based motion estimation for synthetic anima-
tions.
The optical ow eld is a 2D vector eld describ-
ing the translational motion of each pixel from frame
to frame. Our motion estimation algorithm rst com-
|
