Enhanced brain image security using a hybrid of lifting wavelet transform and support vector machine.

Area of Science:

• Computer Science
• Medical Imaging
• Information Security

Background:

• Digital image watermarking is crucial for protecting intellectual property and ensuring data integrity in digital photographs.
• Medical image security is paramount due to sensitive patient data and the need for accurate diagnostics.
• Existing watermarking methods face challenges in balancing robustness against attacks, security, and imperceptibility.

Purpose of the Study:

• To develop an effective and robust digital image watermarking authentication scheme specifically for medical images.
• To enhance the security and resilience of watermarked medical images against unauthorized access and manipulation.
• To ensure high imperceptibility and structural similarity of the watermarked medical images.

Main Methods:

• Utilizing Support Vector Machine (SVM) to differentiate between the Region of Interest (ROI) and Non-Region of Interest (NROI) in medical images.
• Employing Lifting Wavelet Transform (LWT) to embed watermark data securely within the NROI of the cover image.
• Incorporating a shared secret key to bolster the resilience and security of the watermarking scheme.

Main Results:

• The proposed watermarking model demonstrated significant durability and imperceptibility.
• Experimental results yielded a Peak Signal-to-Noise Ratio (PSNR) of 67.81 dB.
• A Structural Similarity Index Measure (SSIM) value of 0.9999 was achieved, indicating excellent image fidelity.

Conclusions:

• The developed SVM and LWT-based watermarking approach provides a robust and secure solution for medical image authentication.
• The scheme effectively preserves image quality while offering strong protection against unauthorized use.
• The method's performance validates its suitability for safeguarding sensitive medical image data.