ABSTRACT
This chapter discusses a new watermarking algorithm using a Rotation, scale, and translation (RST)-invariant feature of images. It begins with a closer look at various geometric distortions and their implications in watermarking methods in “Watermarking Systems and Geometric Distortions”. The chapter explores a new watermarking method based on an invariant feature vector in “A RST-Invariant Watermarking Method” and focuses on two-dimensional images. It describes various geometric distortions that should be considered during watermark design. The most common geometric distortions are rotation, scale, and translation and they belong to rigid-body transformation in that objects in the image retain their relative shape and size. Spatial and spectral domain understanding of the geometric distortions will be applied during the feature vector design and invariance implementation. The watermark is embedded by selecting a vector from the set of extracted feature vectors. Various embedding techniques and capacity issues for the generic feature-based watermark system should be further researched.
