Distinctive Data Hiding in Splines using Robust Image Watermarking

Abstract

This paper presents a robust image watermarking scheme based on a sample spline approach by extracting distinctive invariant features from images that can be used to perform reliable matching between different objects. The features are invariant to basic image transformations like rotation and scaling. The features are highly distinctive, means that every single feature can be correctly matched with high probability against the features of image. This paper also describes an approach called object recognition by using these features. The recognition can be done by matching individual features to a database of features from known objects using a fast nearest-neighbour algorithm. We use the low frequency components of image blocks for data hiding In our watermarking algorithm, to obtain high robustness. We use four samples of the approximation coefficients of the image blocks to construct a spline curve in the 2-D space. The slopes of this spline curve are employed for watermarking purpose. We embed the watermarking code by constructing a spline curve according to watermarking bits. To get a maximum likelihood decoder, we use the distribution of the slope of the embedding spline curve for Gaussian samples.

Keywords: Image watermarking, Maximum Likelihood detector, Hermite Spline, Bessel interpolator, gain attack, invariant features, scale invariance, Steganography, Image, Stego Image.