An effective digital image watermarking scheme incorporating DCT, DFT and SVD transformations

Area of Science:

• Digital Image Processing
• Information Security
• Computer Vision

Background:

• Copyright infringement is a significant concern for digital content owners.
• Existing watermarking techniques offer varying degrees of resistance to different types of attacks.
• Discrete Cosine Transform (DCT), Discrete Fourier Transform (DFT), and Singular Value Decomposition (SVD) possess unique properties beneficial for watermarking.

Purpose of the Study:

• To develop a novel non-blind image watermarking algorithm.
• To leverage the complementary strengths of DFT, DCT, and SVD for robust watermark embedding.
• To enhance resistance against common image distortions and attacks.

Main Methods:

• The algorithm utilizes Discrete Fourier Transform (DFT) for initial decomposition.
• Onion Peel Decomposition (OPD) is applied to frequency sub-images.
• Discrete Cosine Transform (DCT) is applied to frequency bands, followed by zigzag ordering.
• Singular Value Decomposition (SVD) is used for embedding watermark information within the cover image's DFT domain.

Main Results:

• The proposed algorithm demonstrates high-quality watermarked image generation.
• It exhibits strong resistance to various attacks, including JPEG compression and geometric distortions.
• Experimental results confirm the algorithm's stability and robustness across different image databases and conditions.

Conclusions:

• The combined DFT, DCT, and SVD approach offers a superior solution for non-blind image watermarking.
• The algorithm effectively balances embedding capacity with robust protection against image manipulation.
• This method provides a stable and reliable approach to safeguarding digital image copyrights.