In this paper, we propose a scheme in which the watermarking is carried out in the spatial domain using robust feature points to aid in its embedding and extraction. Specific feature points are not only used to embed the information, but also act as a geometrical reference. We propose here a new algorithm that uses the Scale Invariant Feature Transform (SIFT), Shape Theory, and Projective Area-Invariants. The Projective Area-Invariants Transformation allows an image with a distinguished triangle to be put in correspondence with a an affine invariant universal image. The new algorithm is based on embedding the watermark in region around robust feature points and using shape theory to remediate the effects of affine attacks