Artificial Protozoa Optimizer for Enhanced Robust and Secure Image Watermarking

Digital image watermarking faces significant challenges in balancing imperceptibility and robustness under diverse distortions. With the growing prevalence of deepfake and image manipulation technologies, preserving the authenticity and integrity of digital images has become increasingly critical. This paper presents a novel hybrid optimization-based framework for digital image watermarking, integrating the Arnold Transform (AT), Discrete Wavelet Transform (DWT), Singular Value Decomposition (SVD), and the Artificial Protozoa Optimizer (APO). The watermark, such as a QR code, is first encrypted using the Arnold Transform and embedded into the low-frequency sub-band via SVD, while the APO determines the optimal scaling factor α through an objective function combining the Structural Similarity Index Measure (SSIM) and Normalized Cross-Correlation (NCC) ensuring an optimal balance between imperceptibility and robustness. Experiments on a diverse set of images, including medical (e.g., MRI and chest X-ray) and standard benchmark images (e.g., Baboon, Peppers), demonstrate that the proposed framework achieves high SSIM and NCC, along with low Learned Perceptual Image Patch Similarity (LPIPS) and Bit Error Rate (BER). These metrics ensure accurate watermark extraction under attacks such as JPEG compression, noise, and rotation, while supporting ownership protection and multimedia authentication.