Development of Watermarking and Visual Cryptography Based Techniques for Enhancing Digital Image Security

Abstract

The exponential growth in the use of digital images across diverse domains such as communication, healthcare, surveillance, and remote sensing has made image security a critical research concern. The authenticity, confidentiality, and integrity of image data are often threatened by unauthorized access, tampering, and piracy. Conventional cryptographic techniques provide strong protection during storage and transmission, but once the data is decrypted, it remains vulnerable to modification and unauthorized manipulation. Similarly, information hiding techniques such as steganography and robust watermarking ensure ownership protection and covert communication but often fail to provide reliable tamper detection or precise localization. In applications such as satellite imagery, where even minor alterations can have serious implications for decision-making in defense, environmental monitoring, and disaster management, there is an urgent need for fragile watermarking and visual cryptography schemes that balance imperceptibility, accuracy, and robustness. This thesis addresses these challenges by presenting two complementary frameworks—one based on fragile watermarking and the other on visual cryptography—designed specifically to strengthen image authentication and secure communication. The first contribution is a blind fragile watermarking scheme that focuses on tamper detection and localization in satellite images. The scheme introduces a unique watermark generation process where dynamically evolved pixel progenies are used to produce authentication bits, which are then embedded into non-ROI (Region of Interest) regions. A tightly coupled bit-wise relationship matrix and two-level symmetric keys further enhance security. The method achieves high detection accuracy, with tamper localization rates exceeding 93%, while maintaining visual quality with negligible distortion. Experimental analysis demonstrates its effectiveness under various tampering scenarios, including text addition, cropping, and content replacement, confirming its suitability for real-world applications in satellite image integrity verification. The second contribution is a novel 3-out-of-3 visual cryptography scheme for grayscale satellite images that integrates multiple security layers, including image encryption, integrity verification, share meaningfulness, and lossless reconstruction. In this scheme, meaningless shares are avoided by employing Newton basis polynomials for share generation and embedding them into meaningful natural images through the SPICE embedding technique. The resulting shares are not only visually coherent but also carry the authentication information required for verification. During reconstruction, the scheme achieves 100% recovery of the secret image without any contrast loss, with average PSNR values of 46–48 dB, ensuring both imperceptibility and robustness. Furthermore, the scheme resolves major limitations of traditional VC methods by eliminating pixel expansion, preserving share independence, and enabling tamper-proof authentication without additional computation at the receiver’s end. Extensive experimental validation against state-of-the-art fragile watermarking and visual cryptography techniques demonstrates that the proposed approaches outperform existing solutions in terms of imperceptibility, authentication accuracy, and robustness against attacks. The fragile watermarking framework excels in precise tamper localization and low computational cost, while the visual cryptography framework ensures confidentiality, authentication, and integrity verification in a single process. Together, these contributions form a comprehensive security model for digital images, with strong relevance to domains such as satellite communication, defense image analysis, copyright enforcement, and secure transmission of sensitive multimedia data. By integrating fragile watermarking and visual cryptography, this research advances the state-of-the-art in image security, providing scalable, efficient, and practical solutions that bridge the gap between theoretical cryptographic strength and real-world applicability.