Wavelet-Assisted Steganography and Visual Secret Sharing for Privacy-Preserving Medical Image Exchange

Abstract

Medical exchange of images should be secure to ensure patient privacy and not compromise diagnostic quality. The paper describes a wavelet-based steganography and secret sharing scheme of medical image sharing to ensure privacy. In the first step medical images are pre-processed to make them clear and free of noise without compromising the diagnostic value. The wavelet-based steganography is used to embed sensitive patient metadata into the medical image, resulting in stego images that are visually identical to the originals. In order to increase the security, the stego images are encrypted and broken into various random blurred shares by the use of visual secret sharing methodologies such that the individual shares do not reflect any significant information. The reconstruction process is authorized where the required shares are combined and the original medical image and metadata are restored. The experimental evidence proves the proposed approach to provide satisfactory image quality regarding PSNR 15.7-16.0 dB and SSIM 0.847-0.857, which indicate that there is a structural similarity between the original and stego images. The analysis of histogram demonstrates that there is little statistical disparity and the system is resistant to steganalysis attacks. The system is found to be reliable in reconstruction with 100% accuracy in recovery and classification accuracy is found to be approximately 58 which shows the strength and practicability of the method. All in all, the suggested framework is a useful solution to the issue of safe and privacy-saving medical image exchange within a healthcare setting.