Robust zero-watermarking method for multiple medical images using wavelet fusion and DTCWT-QR

Abstract

Following smart healthcare’s radical progression, medical images face security concerns, such as information leakage and malicious tampering. To enable copyright protection and cost reduction for multiple medical images, we propose a zero-watermarking method for multiple medical images with joint dual-tree complex wavelet transform (DTCWT), QR decomposition, and discrete cosine transform (DCT). First, numerous medical images are fused using a wavelet-transform-based fusion method to reduce costs. Next, the generated image is subjected to the DTCWT, which divides the resulting low-frequency sub-bands into blocks, after which each block is subjected to DCT and QR decomposition, generating a significant binary image using the magnitude of the relationship between the value of the two-norm of the first row of each block and the overall mean value. Lastly, a method based on improved Hénon mapping image encryption is presented for encrypting a copyrighted image. An XOR operation is performed on the encrypted copyrighted image with a binary image to yield a zero-watermarking image. Numerous results show that our method can withstand different types of attacks with the normalized correlation coefficients remaining higher than 0.96. Moreover, the proposed method achieves superior robustness with an average performance improvement of approximately 3.3% compared with the latest and similar methods. These results demonstrate the superiority of our method, which can be applied in copyright protection applications for medical images.