A novel blind color image watermarking based on Walsh Hadamard Transform

• Published in: Multimedia tools and applications 2020-03, Vol.79 (9-10), p.6845-6869
• Main Authors: Prabha, K., Sam, I. Shatheesh
• Format: Article
• Language: English
• Subjects: Algorithms; Coefficients; Color; Color imagery; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Decomposition; Digital broadcasting; Digital media; Digital watermarks; Embedding; Encryption; Multimedia; Multimedia Information Systems; Signal to noise ratio; Special Purpose and Application-Based Systems; Spread spectrum; Watermarking; Wavelet transforms
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-019-08212-w

A new blind color image watermarking based on Walsh Hadamard Transform (WHT) is proposed in this paper which uses Triangular Vertex Transform (TVT) along with WHT. This method introduces a Triangular Vertex Transform (TVT) and Inverse Triangular Vertex Transform (ITVT) as well as applies this transform for image watermarking. Initially, convert the R, G, B color image to TVT coefficients U, V and W using the proposed TVT transform. Then the W coefficient is subdivided into 4 × 4 non-overlapping blocks, which are then transformed using WHT. The watermark is encrypted with the key K using the 2D-Logistic Sine Coupling Map encryption to obtain the encrypted watermarked image. The bits of the encrypted image is embedded in the first row WHT coefficients since the embedding of watermark bit on the first row provides high robustness to attacks. After embedding the data, the inverse WHT and inverse TVT are applied to obtain the watermarked image. The extraction process is similar to that of the embedding process where TVT is applied on the Watermarked image to construct W′ coefficient. Subdivide the attained W′ coefficient into 4 × 4 sub-block and then the WHT is applied to the 4 × 4 sub-block for extracting the data from the first row of WHT coefficients. The performance of the proposed watermarking algorithm outperforms in terms of Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSIM) and Normalized cross-correlation (NC). Experimental results reveal that the proposed watermarking method is highly resistant to different types of attacks.