Rotation, scaling and translation resilient digital image watermarking using log-polar mapping

Digital image watermarking or data hiding techniques must resist various attacks such as compression, noising and geometric distortion attacks. Rotation, scaling and translation (RST) in geometric distortion attacks are considered to be more challenging than other attacks. The logpolar mapping (LPM) coordinate is used to analyse the angle of rotation and ratio of scaling between the cover image and the corrupted watermarked image. However, the watermark or secret data are easily destroyed when inverse LPMis applied. Based on LPM, this paper presents a robust digital image watermarking technique. We discuss the appropriate watermark embedding locations in the LPMcoordinates. The experiment results show that the proposed method has high wPSNR and low bit error rate (BER), and can withstand RST attacks.