Integrating DCCS-Net for High-Fidelity Watermarking and Tampering Detection in Brain Scan Images
BSH. Shayeez Ahamed, Radhika Baskar, Ganapathi Nalinipriya · 2024
This study presents a new approach that combines the Deep Cascade Compressed Sensing Network (DCCS-Net) with a watermarking and tampering detection framework to improve the accuracy and quality of medical images. The system analyzes brain scan images by incorporating watermarks and identifying any tampering, thereby guaranteeing a high level of accuracy and maintaining the structural resemblance to the original images. The essential steps involve preprocessing the image using Discrete Cosine Transform (DCT), employing compressed sensing for efficient data management, and embedding watermarks using binary quantified reference values. The evaluation metrics, namely Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM), indicate that the watermarked images exhibit a PSNR of 41.906 and SSIM of 0.98458. These values suggest that there is minimal degradation in the quality of the images. The tampering detection process achieves a precision of 0.94531, effectively identifying tampered blocks. The reconstructed images exhibit minimal degradation in quality, as evidenced by a PSNR of 36.21 and SSIM of 0.9789. The results highlight the ability of DCCS-Net to effectively manage noise, minimize artifacts, and maintain intricate details, thereby improving the accuracy and reliability of diagnostics. The graphical depiction of metrics provides additional evidence to support the strength and reliability of the methodology in preserving image quality and reconstructing manipulated images, thus ensuring the trustworthiness and reliability of medical image analysis.