Lossless fragile pinpoint authentication scheme for medical images

This paper proposes an imperceptible, reversible watermarking scheme based on messy system for the authentication of medical images. The digital fundus images are one particular class of medical images which has been chosen for simulation and analysis of the proposed scheme in this paper. These images were given in TIF format in RGB color space. The proposed scheme dynamically generates the watermark using messy models. And, it is embedded inside the image by expanding intra plane difference between any two color planes of images. It is known as intra-plane difference expanding. The experimental results shown the improvement in the imperceptibility by 15% and 30000 bits were found as the maximum possible size of the watermark with good level of the imperceptibility which is above 60dB at an average. Moreover, the proposed scheme maintained the detection accuracy rate at 100% without any loss in the process. The proposed scheme is very sensitive to the jittering, geometrical and various filtering attacks; thereby it modifies around 40% of the watermark for jittering and geometrical attacks and around 45% of the watermark for filtering attacks at an average and which are sufficiently large enough to ensure the better security for the authentication of medical images.