A secure watermarking technique to address security concerns of modern telemedicine applications: a chaotic map-based approach

The modern era has witnessed an immense propagation of multimedia applications. Authentication and copyright are hot topics with these next-generation multimedia platforms. Telemedicine is one perspective domain that customs the benefits of multimedia ecosystems to deliver efficient medical services. In addition, the recent COVID-19 pandemic aids as a catalyst to put telemedicine in the spotlight of modern research. The spike in telemedicine usage has led to a rapid escalation in the generation and sharing of electronic medical data. Medical image watermarking is a complementary security mechanism to address the issues of integrity, copyright, and confidentiality. This manuscript presents a robust, imperceptible, and efficient scheme for watermarking medical images to protect legitimate ownership and privacy. The architecture integrates three heterogeneous chaotic maps (Chebyshev, Tangent Delay Ellipse Reflecting Cavity System, and an Extended logistic map) to significantly expand the effective key space and enhance key sensitivity. The resiliency of the algorithm is further enriched by supplementing confusion operation in the watermark using a 4 × 4 Substitution box. Three unique chaotic maps are used to amplify the endurance of the proposed scheme, which specifies the location for watermark insertion The chaotic sequences mutually govern the spatial and frame-wise watermark embedding locations. Experimental evaluations conducted on medical images demonstrate excellent visual imperceptibility by having peak signal-to-noise ratio (PSNR) values exceeding 55 dB and structural similarity index (SSIM) values close to unity. Statistical integrity is validated by Chi-Square analysis (X² = 18,695,731), confirming histogram preservation. Security is bolstered by an extensive key space of 10¹⁰⁸, equivalent to 256-bit protection, while a Crest Factor of 5.0989 ensures a noise-like encrypted distribution. Robustness analysis under noise addition and cropping, yields low bit error rate values, validating reliable blind watermark extraction without the original carrier image. Results show that the scheme has excellent attributes to handle authentication and copyright issues for real-time medical applications.