Data throughput and robustness studies in a new digital image watermarking technique based upon adaptive segmentation and space-frequency representation (WASSFR)

This paper presents enhancements in the form of improved segmentation and transforms for the novel watermarking technique based on adaptive segmentation and space-frequency representation we developed earlier (Sinha, 2000). We also provide details of the robustness studies that have been performed for this novel watermarking technique. Furthermore, we describe a complete digital rights management system including synchronization and channel coding suitable for practical applications such as secure digital image distribution. The watermarking technique proposed in Sinha is based upon adaptive selection of a suitable space-frequency representation for each segment of the image. Suitable segmentation of the image for this purpose is obtained by using a new technique described there. The watermark is then added to the perceptually significant components in each of the transformed segments. The proposed approach allows for the incorporation of a significantly stronger watermark and higher data throughput when compared to conventional frequency domain watermarking techniques).