Novel Digital Image Watermarking Algorithm Combining Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD)

A novel digital image watermarking algorithm combining Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) is proposed. The algorithm aims to protect digital images while embedding watermark information for copyright protection and authentication purposes. Implementation begins with wavelet decomposition of the original image using DWT, where the image is transformed into frequency subbands. The low-frequency subband (LL subband) is then partitioned into smaller blocks for localized processing. Each block undergoes SVD decomposition, which factorizes the block matrix into three components: U, S, and V matrices. The diagonal S matrix contains singular values representing the block's energy distribution. For watermark embedding, the largest singular value from each block is selected and modified using quantization-based embedding to incorporate the watermark data. The watermark is preprocessed using Arnold scrambling transformation to enhance security, which cyclically permutes pixel positions through iterative coordinate transformations. The algorithm maintains image integrity by leveraging DWT's multi-resolution properties and SVD's stability characteristics. Watermark extraction follows the reverse process: Arnold descrambling, singular value extraction from decomposed blocks, and quantization-based detection. This approach effectively balances robustness and imperceptibility while supporting reliable copyright verification for digital content protection.