Analysis and computation of projective invariants from multiple views in the geometric algebra framework

A central task of computer vision is to automatically recognize objects in real-world scenes. The parameters defining image and object spaces can vary due to lighting conditions, camera calibration and viewing positions. It i s therefore desirable to look for geometric properties of the object which remain invariant under such changes. In this paper we present geometric alg ebra as a complete framework for the theory and computation of projective i nvariants formed from points and lines in computer vision. We will look at the formation of 3D projective invariants from multiple images, show how th ey can be formed from image coordinates and estimated tensors (F, fundament al matrix and T, trilinear tensor) and give results on simulated and real d ata.