Research On Geometrically Resilient Medical Image Robust Zero-Watermarking Algorithms

With the rapid development of information technology,digital medical image is applied increasingly in disease diagnosis and treatment.Nevertheless,while enjoying the convenience of medical images,there are risks of illegally copying and tampering of medical image data without authorization through public network transmission.Therefore,it is particularly critical to protect the copyright ownership of medical images,in which robust zero-watermarking is very suitable for medical images out of its adoption of zero-embedding method,which does not modify the original image,enabling copyright protection with guaranteed visual quality and integrity of the image.The research of robust medical image zero-watermarking mainly focuses on the ability of the algorithm to resist various attacks,that is,the designed algorithm can completely extract the zero-watermark information after various attacks.Because the geometric attack will change the position relationship between the attacked medical image and the original one,which will cause a serious desynchronization between the zero-watermark information and the medical image,making the extraction of the zero-watermark information extremely difficult.The ability of the algorithm to resist geometric attacks has become a research hotspot in medical image zero-watermarking,and it is also the core issue of this dissertation.In this dissertation,four zero-watermarking algorithms for medical images are designed based on the geometric invariance of quaternion continuous orthogonal moments.The main work of this dissertation is as follows:(1)A new robust geometrically resilient zero-watermarking algorithm for color medical images is proposed based on quaternion polar harmonic transforms(QPHTs)and mixed linear-nonlinear coupled map lattice(MLNCML).Firstly,the precise coefficients selection of QPHTs is discussed,followed by the design of zero-watermark construction and zero-watermark detection method on the basis of QPH I s.In the procedure of zero-watermark construction,magnitudes of the precise QPHTs are utilized to construct zero-watermark,which ensures the robustness of the algorithm against geometric attacks;furthermore,MLNCML system is adopted to scramble the constructed feature images to improve the security of the algorithm.The experimental results showed that the algorithm has good robustness against common image processing attacks,image rotation and image scaling,which is superior to similar zero-watermarking algorithms.(2)To improve the accuracy of QPHFMs,QPHFMs computation method with high-precision was proposed via adopting wavelet numerical integration(WNI),which promoted the accuracy effectively.Then,a new geometrically resilient robust zero-watermarking algorithm for color medical images is proposed on the basis of the proposed precise QPHFMs and chaotic maps.The algorithm applied magnitudes of precise QPHFMs to construct feature image,the good geometric invariance of which provided the proposed zero-watermarking algorithm with good anti-geometric attack ability.The experimental results demonstrated that the algorithm is robust to common image processing attacks,as well as geometric attacks such as scaling,rotation,length-width ratio changing and flipping.Compared with other zero-watermarking algorithms and other color medical image watermarking algorithms,the performance of this algorithm is superior.(3)A multiple color medical image zero-watermarking algorithm against complicated geometric attacks is proposed with the combination of the image local stability feature region and QPHFMs.This algorithm utilized the SURF operator to extract stable feature points from the original color medical image.Then,the locally stable feature regions were determined auto-adaptively according to the feature scales,and the local feature regions were standardized.Finally,the QPHFMs of the local feature regions were calculated,and several feature vectors were constructed by their magnitudes,which were binarized to form multiple zero-watermarks.The trial revealed that the proposed multiple zero-watermarking copyright authentication algorithm can not only resist common image processing attacks and general geometric attacks,but also have good robustness against complicated geometric attacks,such as translation,cropping,length-width ratio changing and Unzign.Compared with the embedded watermarking algorithm based on local feature region,this algorithm performs better.(4)The existing zero-watermarking algorithms are mainly for copyright protection of one image,which can be used for multiple images repeatedly,whereas it takes a lot of time and occupies a big amount of storage space during repeated operations.To solve this problem,a novel geometrically resilient zero-watermarking algorithm is proposed,which can simultaneously protect the copyright of three medical images via the application of QPHFMs and taking three medical images as a whole.Firstly,three medical images were taken as three imaginary components of a set of pure quaternions,and the QPHFMs of the three images were computed.Then,the magnitudes of QPHFMs were adopted to construct the feature image,and XOR operation was carried out on the scrambled feature image and Logo image to obtain zero-watermark image.The algorithm can achieve simultaneously copyright protection of three images,which could improve the efficiency of zero-watermarking system and save storage space as well.The experimental results showed that the algorithm can effectively resist common image processing attacks as well as geometric attacks,which can be well applied to copyright protection of three images.