云环境下基于密文域的医学图像鲁棒水印算法研究

发布时间：2018-06-13 12:02

本文选题：密文域 + 加密医学图像　；参考：《海南大学》2017年硕士论文

【摘要】：为了保护个人信息,前人在数字水印领域进行了大量的研究。然而,以往的数字水印算法,几乎都是在明文域(即未加密的载体图像)中进行水印的嵌入和提取,只能保护水印信息的安全,忽略了对载体图像安全性的考虑。对于医学图像这一类特殊类型的图像,因图像本身包含着大量的重要信息,也属于需要保护的对象。明文域下的医学图像水印算法不能满足上述需求,需要在密文域下寻找合适的解决途径。在图像加密算法方面,前人取得了很多研究成果,但在密文域下进行鲁棒数字水印时效果并不理想,并不是所有的图像加密算法都能保证密文域下的数字水印鲁棒性要求;利用全同态加密技术可以实现上述操作,但运算时间开销巨大,不能满足实际需求。考虑以上几点原因,我们在以下几个方面开展了密文域医学图像鲁棒数字水印算法的研究。研究了云环境下基于DCT的加密医学图像鲁棒水印算法。通过DCT变换提取加密医学图像的特征向量,将水印序列和其进行关联,在密文域下实现水印的嵌入和提取。测试数据表明,提出的算法不仅对JPEG压缩、高斯噪声攻击、中值滤波、缩放、旋转、平移等图像攻击具有较好的鲁棒性,还具有理想的加密算法同态性,可同时保护原始医学图像和水印信息。研究了云环境下基于DFT的加密医学图像鲁棒零水印算法。使用DFT变换和Logistic混沌序列将原始医学图像和水印同时加密。提取加密医学图像的DFT中低频系数构造特征向量。采用零水印技术,将水印信息与特征向量进行关联,在加密图像中完成水印的嵌入和提取。实验结果表明,密文域算法除具有很好的抗常规攻击和几何攻击鲁棒性外,还具有理想的加密算法同态特性。研究了云环境下基于DWT-DCT和DWT-DFT的加密医学图像鲁棒零水印算法,可以在不修改图像像素值的前提下完成密文域水印的嵌入和提取,保证了医学图像的完整性。实验结果表明,提出的算法不仅具有良好的抗常规攻击和几何攻击鲁棒性,还具有理想的加密同态性和运算速度。借助加密算法的同态性,可将密文状态下的载体医学图像和水印交给第三方,在云环境下实现水印的嵌入和提取。
[Abstract]:In order to protect personal information, many researches have been done in the field of digital watermarking. However, most of the previous digital watermarking algorithms are embedded and extracted in plaintext domain (i.e. unencrypted carrier image), which can only protect the security of watermark information and ignore the security of carrier image. For the medical image, which is a special type of image, because the image itself contains a lot of important information, it also belongs to the object in need of protection. The medical image watermarking algorithm in plaintext domain can not meet the above requirements, so it is necessary to find a suitable solution in ciphertext domain. In the aspect of image encryption algorithms, many achievements have been made, but the effect of robust digital watermarking in ciphertext domain is not ideal, and not all image encryption algorithms can guarantee the robustness of digital watermarking in ciphertext domain. The operation can be realized by using the full homomorphism encryption technology, but the computation time is too expensive to meet the actual demand. Considering the above reasons, we study the robust digital watermarking algorithm of medical image in ciphertext domain in the following aspects. A robust watermarking algorithm for encrypted medical images based on DCT in cloud environment is studied. The feature vector of encrypted medical image is extracted by DCT transform, and the watermark sequence is associated with it. The watermark is embedded and extracted in ciphertext domain. The test data show that the proposed algorithm is not only robust to JPEGcompression, Gao Si noise attack, median filtering, scaling, rotation, translation and other image attacks, but also has the ideal homomorphism of encryption algorithm. The original medical image and watermark information can be protected simultaneously. A robust zero watermarking algorithm for encrypted medical images based on DFT in cloud environment is studied. DFT transform and Logistic chaotic sequence are used to encrypt the original medical image and watermark simultaneously. The feature vectors are constructed by extracting the low frequency coefficients of the encrypted medical images in DFT. The watermark is embedded and extracted from the encrypted image by using the zero-watermark technique, which correlates the watermark information with the feature vector. Experimental results show that the ciphertext domain algorithm not only has good robustness against conventional and geometric attacks, but also has the ideal homomorphism of encryption algorithm. A robust zero-watermarking algorithm for encrypted medical images based on DWT-DCT and DWT-DFT in cloud environment is studied. The watermark embedding and extracting in ciphertext domain can be completed without modifying the pixel value of image, and the integrity of medical image can be ensured. Experimental results show that the proposed algorithm not only has good robustness against conventional and geometric attacks, but also has ideal encryption homomorphism and computational speed. With the help of the homomorphism of the encryption algorithm, the carrier medical image and watermark in the ciphertext state can be handed over to a third party, and the watermark can be embedded and extracted in the cloud environment.
【学位授予单位】：海南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP309.7

【相似文献】