Implementation of an Efficient Secure Hybrid Reversible Tampering Detection Method in Medical Image

Abstract

Medical image watermarking has gained significant attention in recent years as a critical tool for data security. Given that even minor distortions can compromise diagnostic accuracy and lead to misinterpretation, reversibility is paramount in medical imagery. This research introduces a sophisticated hybrid reversible watermarking architecture operating within the frequency domain. By leveraging the combined strengths of Discrete Wavelet Transform (DWT), Discrete Cosine Transform (DCT), and Singular Value Decomposition (SVD), the system achieves an optimal balance between payload capacity, imperceptibility, and robustness. The proposed method generates the watermark directly from the host image's own features to enable self-embedding, ensuring highly accurate blind tamper detection and integrity verification without requiring the original image. To address security and authentication concerns, a dual watermarking strategy is implemented. A reversible watermark (W1), generated through the RLSB method, is embedded into the LL coefficients to preserve diagnostic integrity, while a secondary biometric logo (W2) is integrated into the HH coefficients to facilitate authentication and confidentiality. Experimental results demonstrate high robustness against common signal processing attacks, while comparative analysis using AI-driven classification features confirms the superior performance of the proposed approach over existing methods.

Keywords