Stereo tracking using ICP and normal flow constraint

This paper presents a new approach for 3D view registration of stereo images. We introduce a hybrid error function which combines constraints from the ICP (iterative closest point) algorithm and normal flow constraint. This new technique is more precise for small movements and noisy depth than ICP alone, and more robust for large movements than the normal flow constraint alone. Finally, we present experiments which test the accuracy of our approach on sequences of real and synthetic stereo images.