КРИТЕРІЙ РОБАСТНОСТІ У ЦИФРОВИХ ВОДЯНИХ ЗНАКАХ НА ЗОБРАЖЕННІ

In the context of increasing circulation of digital visual content, ensuring authenticity, copyright protection, and resilience of images to tampering has become a significant challenge. Digital watermarking represents one of the key methods for protecting visual information by embedding auxiliary data into the image in a manner that is imperceptible to the human eye but detectable by specialized algorithms. However, practical application of watermarking critically depends on robustness — the ability of the embedded watermark to withstand various forms of intentional or unintentional image processing, including compression, noise addition, geometric transformations, and filtering. This article provides a comprehensive analysis of the robustness criterion in digital image watermarking systems. The study systematizes common categories of attacks (e.g., JPEG compression, Gaussian noise, rotation, cropping, scaling) and investigates how these affect the integrity of watermarks. A structured review of objective evaluation metrics is presented, focusing on Bit Error Rate (BER), Normalized Cross-Correlation (NCC), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM). These indicators allow for a quantitative assessment of both detection reliability and visual quality preservation after watermark embedding. The work also explores the inherent trade-offs between robustness and other key parameters such as imperceptibility, payload capacity, and security. Through analysis of real-world scenarios, it demonstrates how reinforcing one aspect often negatively impacts the others — for instance, higher robustness may lower image quality or reduce embedding capacity. The article highlights the importance of achieving a balanced design approach tailored to specific application goals — whether for copyright enforcement, image authentication, or metadata embedding. Adaptive strategies, such as human visual system (HVS) masking and the integration of cryptographic methods, are proposed as effective tools for optimizing watermark embedding. Future research directions include the development of machine-learning-based adaptive embedding schemes, context-sensitive visual masking, and robust cryptographic coding. These advancements aim to enhance the reliability, flexibility, and practical utility of watermarking systems in diverse digital imaging environments.