MEDICAL IMAGE AUTHENTICATION WITH ENHANCED WATERMARKING TECHHNIQUE THROUGH VISUAL CRYPTOGRAPHY

Technological evolution has increased the insecurity of the biomedical images produced day to day. Avoidance of the database hacking and preserving the privacy of digital biometric data stored in the central database has become of paramount importance. In order to preserve the security of the database, an enhanced visual cryptographic technique is proposed in this paper. Cryptographic technique allows information to be encrypted in such a manner that decryption cannot be handled by the non-legitimate user. The proposed method allows a secret sharing scheme, where an image is broken up into two shares and then overlaid in such a way that an unauthenticated person cannot reveal this secret image. The partitioned image is compared with all the images in the target using Hausdorff Distance metric algorithm. When both the images match, the authentication is provided. This paper also focuses on outsourcing data to an untrusted client to predict the category of the image without getting any information about the image itself and the final result of the classification. The experimental results show that our proposed method is efficient over the existing methods.