This paper addresses feature-based image watermarking methods for
digital contents. Most previous watermarking algorithms suffer from geometric
distortion attacks that desynchronize the location of the inserted watermark. In
order to synchronize the watermark location, watermark synchronization should
be performed. One solution for watermark synchronization is to use image features.
This paper describes a feature-based watermarking method based on
scale-invariant keypoints. We extract feature points from the scale-invariant
keypoint extractor and then decompose feature points into a set of disjoint triangles.
These triangles are watermarked by an additive spread-spectrum method
on the spatial domain. We perform an intensive simulation in comparison with
other feature-based watermarking methods using 100 test images. Results show
that the proposed method is considerably robust against both geometric distortion
attacks and signal processing attacks listed in Stirmark 3.1.