Conformal Geometric Algebra for Motion and Shape Description

Applications of conformal geometric algebra (CGA) in problems of computer vision and graphics related to motion and shape description show that,CGA can provide universal and effective representations and algorithms.By now,such applications have been focused on adopting the Grassmannian structure, the spinor and twist representations of rigid body motions.
By analyzing different representations of rigid body motions,we show how the Grassmannian structure is applied to monocular vision problems and simplifies the solving procedure.We the n show how the spinor and twist representations can reduce the number of nonlinear constraints in the space of parameters,in sharp contrast to the matrix representation of rigid body motions.This feature makes the spinor and twist representations often very effective in solving problems like pose estimation,shape approximation and curve blending,to name a few.