Hybrid medical image zero watermarking via discrete wavelet transform-ResNet101 and discrete cosine transform

Abstract

This paper suggests a way to move forward with medical images based on the discrete wavelet transform (DWT), ResNet101-DCT (discrete cosine transform), and zero watermarking. In addition, we extract deep medical image attributes using ResNet101, a pre-training network. Furthermore, deep features undergo a transformation via DWT and DCT, then a perceptual hash function is employed to generate the feature vector. The watermark is encrypted by chaotic scrambling the original watermark, an exclusive-OR (XOR) gate into the medical image, and generating and storing the logical key vector. Overall, the feature extraction technique is employed to extract the significant characteristics of the medical image under testing and develop the feature vector. Finally, the feature vector is subjected to an XOR maneuver with the logical key vector, deriving the encrypted watermark. Once the encrypted watermark has been obtained and decrypted, the restored watermark can be used to determine the medical image's ownership and watermark information by calculating the normalized correlation (NC) coefficient. Most NC values are greater than 0.50. The algorithm also shows good robustness against conventional and geometric attacks.