MATLAB Image Digital Watermarking Using DCT and DWT Algorithms

In this graduation thesis, I present a MATLAB-based digital image watermarking scheme that combines Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) algorithms. The thesis provides an in-depth explanation of the hybrid algorithm and includes complete program code implementation. This approach leverages the complementary characteristics of DCT's energy compaction properties and DWT's multi-resolution analysis to create an effective digital watermarking method for images. The implementation involves several key MATLAB functions including dct2() for 2D discrete cosine transform, wavedec2() for 2D wavelet decomposition, and idct2() along with waverec2() for the inverse transformations. The algorithm workflow typically includes: preprocessing the watermark image, selecting appropriate DCT coefficients from image blocks, embedding watermark information into specific DWT subbands (particularly the HL and LH subbands for better robustness), and applying inverse transforms to reconstruct the watermarked image. By integrating modern watermarking techniques with traditional image processing algorithms, this thesis demonstrates how to achieve improved watermark protection and enhanced image security. The code implementation includes parameters for adjusting watermark strength, controlling the trade-off between visibility and robustness, and handling different image formats. Additionally, the thesis provides comprehensive analysis and discussion of experimental results, including performance metrics such as Peak Signal-to-Noise Ratio (PSNR) and Normalized Correlation (NC) to validate the scheme's effectiveness and feasibility against various attacks like compression, filtering, and noise addition. Through this work, readers will gain deep understanding of MATLAB-based digital image watermarking techniques and can practically implement the described methods using the provided program code for further research and application development. The code structure includes modular functions for embedding, extraction, and evaluation, making it suitable for both educational purposes and advanced research in digital watermarking systems.