Watermarking in the space/spatial-frequency domain using two-dimensional Radon-Wigner distribution

Digital watermarking remains a key challenge in secure media distribution. Existing methods often struggle with robustness against common image manipulations. Prior research has shown that spatial domain techniques can be vulnerable to attacks like compression or filtering. No prior work had resolved the issue of embedding watermarks that remain detectable after such transformations. This gap motivated the development of new watermarking strategies. Researchers have explored frequency domain methods for better resilience. However, these approaches may not capture both spatial and frequency characteristics effectively. The need for a dual-domain solution became evident. This paper introduces a novel approach using 2-D Radon-Wigner distributions.

This study aims to enhance watermarking robustness by leveraging both spatial and frequency domains. The authors propose a watermark with variable spatial frequency. This design allows for better adaptability to image transformations. The goal is to create a watermark that remains detectable after common attacks. The study focuses on linear frequency changes in 2-D signals. It tests the effectiveness of 2-D space/spatial-frequency distributions. The researchers aim to demonstrate resilience against translation and rotation. They also evaluate performance under JPEG compression and filtering.

The researchers designed a 2-D watermark with linear frequency variation. They embedded this watermark in the spatial domain of images. The detection process uses 2-D space/spatial-frequency distributions. The 2-D Radon-Wigner distribution is applied to emphasize detection. This method projects the Wigner distribution to highlight watermark features. The study tests watermark robustness using Stirmark 3.1. Several image processing attacks are simulated in the experiments. The researchers measure detection accuracy after each transformation.

The proposed watermark showed strong resilience to translation and rotation. It remained detectable even after significant image cropping. JPEG compression at 80% quality did not eliminate the watermark. Filtering operations had minimal impact on watermark visibility. The 2-D Radon-Wigner distribution improved detection accuracy. The watermark maintained a detection rate above 90% in most cases. Performance was consistent across different attack combinations. These results suggest the method is suitable for practical applications.

The authors propose that the 2-D Radon-Wigner distribution enhances watermark detection. They suggest that linear frequency changes improve robustness against attacks. The study supports the claim that this method outperforms traditional approaches. The results indicate strong performance under JPEG compression and filtering. The researchers propose that this method is suitable for secure media distribution. They suggest that the 2-D Radon-Wigner distribution is a valuable tool. The study supports the claim that this approach maintains watermark integrity. The authors suggest that this method may be adapted for other signal types.