FRRW: A feature extraction-based robust and reversible watermarking scheme utilizing zernike moments and histogram shifting
Abstract
This paper introduces a feature extraction-based approach to ensure both robustness and reversibility of image. Low-order Zernike moments are utilized to embed a robust binary image as a watermark, which is used for information authentication. A reversible watermark is embedded outside the robust watermark regions and is employed for the purpose of restoring the cover image. It uses the combination of histogram shifting and prediction error, which can improve image restoration quality. Steady feature points are extracted in two ways, the speed-up robust features (SURF) algorithm and the oriented fast and rotated brief (ORB) algorithm. After extracting the feature points, the regions are obtained by extending the final selected feature points to embed the watermark. Consequently, the presented watermarking technique combines robust and reversible watermarking which has the ability to enhance the invisibility of the watermark and the clarity of image restoration. It is possible to extract the watermark even after an attack has been made on the watermarked image. Or we can recover the original image with no attacks. The results from the experiments indicate that the suggested method is resilient to geometric deformations, involving scaling and rotation, along with typical signal manipulation attacks, including noisebased attacks.
Sun Ying
Phd student
Xiaochen Yuan
Associate Professor
