面向多应用场景的加密图像可逆信息隐藏方法研究

发布时间：2018-03-07 14:20

本文选题：加密图像可逆信息隐藏　切入点：嵌入混乱相结合　出处：《中国科学技术大学》2017年硕士论文　论文类型：学位论文

【摘要】：现如今云存储十分流行,大大节约本地空间。然而数据外包存储也会导致个人企业的隐私泄漏。用户为了保护隐私信息,以图像为例,可以对其"加密"。而云端在接收到密文图像后,为了方便密文图像的管理查询、所属方认证以及保障原始数据真实性,可以嵌入用户ID以及认证标注等水印信息。在向密文图像嵌入水印信息后,云端不能对图像造成永久失真,需要同时保证水印信息的正确提取和密文图像的无损复原。加密图像可逆信息隐藏技术,能够在密文图像中进行水印信息的嵌入,并能确保在水印正确提取后,可以无损地重构出密文图像,同时能正确解密得到明文图像,保障用户数据的完整性、真实性。已有的方法只针对云端嵌入不可视水印展开研究,云端采用的可逆隐藏技术必须与发送方(接收方)采用的加密(解密)技术对应,这在现实场景中不太适用。不仅如此,已有的方法没有考虑到以下场景:(1)云端被认为是"半诚信式"的,用户也需要嵌入水印对数据进行完整性真实性认证;(2)云端需要嵌入可视水印例如公司LOGO、用户信息,从而可以更加轻松地识别出密文数据的所属方。本文全面考虑上述多种场景,对不同场景下的加密图像可逆信息隐藏技术进行研究,内容列举如下:1.基于嵌入混乱相结合的用户端可逆信息隐藏算法论文针对用户端嵌入不可视水印的场景,面对既要加密图像内容,又要实现大嵌入率下的可逆信息隐藏的两个要求,首先精准预测像素降低预测误差的信息量,而后在进行信息嵌入过程的同时混乱图像内容,严重破坏图像内容质量,达到加密目的。论文提出的局部线性预测模板,预测效果精准,使得大量嵌入信息成为可能。2.基于可逆图像变换的云端密文域可逆信息隐藏算法论文针对已有的云端嵌入不可视水印场景下的方法缺陷,提出一种与"用户端无关"的云端密文域可逆信息隐藏算法,具有实际应用价值。该方法以可逆图像变换为基础,将被保护的明文秘密图像以另外一幅明文目标图像为参照进行变换,得到与目标图像内容相近的明文变换图像,秘密图像语义内容被目标图像内容掩盖保护。由于变换后的图像也是明文,所以云服务器端嵌入水印信息的方法可以是任意一种明文可逆信息隐藏算法,该算法独立于用户端发送方和接收方。3.云端密文域可逆的可视水印算法论文针对云端嵌入可视水印的场景,提出一种新颖的密文域可逆的可视水印算法。已有的可逆可视水印算法都是针对明文图像,通过记录原始图像和嵌入水印的载秘图像的差值达到可逆性。由于密文图像像素之间完全没有相关性,导致预测误差仍然是随机分布,产生的大量差值难以被可逆地嵌入到图像中。本文的做法是在原始明文图像中预留部分空间,该空间在加密后用来嵌入可视水印,从而避免差值的记录。预留空间的像素可以采用明文可逆信息隐藏方法嵌入非水印区域中,从而得以无损重构。
[Abstract]:Now cloud storage is very popular, greatly save space. However, local data outsourcing storage will lead to individual enterprises. In order to protect the privacy of user privacy information, to the image as an example, the "encryption". The cloud in the received cipher image, in order to facilitate query encrypted image management, the party certification the original data and guarantee the authenticity, the user can embed ID certification mark and the watermark information. The watermark information is embedded in the image to the ciphertext, the cloud can not cause permanent distortion of the image, at the same time to ensure correct extraction watermark and encrypted image lossless restoration of image encryption. Reversible information hiding technology, can the watermark information in the ciphertext in the image in the watermark embedding, and can ensure the correct extraction, lossless reconstruct the cipher image, and can correctly decrypt the image, protect the user data Integrity and authenticity. The existing methods for cloud embedding visible watermark is studied, using the cloud reversible hiding technology must be with the sender (receiver) using encryption (decryption) corresponding to this technique is not applicable, in the real scene. Moreover, the method does not take into account the following scenarios: (1) the cloud is considered to be the "semi honest", users also need to embed the watermark integrity authentication of data; (2) cloud embedding visible watermark such as LOGO, user information, and can more easily identify encrypted data is. Considering the various scenes, hiding technology research on image encryption reversible information in different scenarios, the contents are listed as follows: 1. based on the end user reversible information hiding algorithm combining chaos embedded according to user end embedding watermark scene, face For both to encrypt the image content, but also to achieve the two requirements of the reversible information hiding high embedding rate, accurate prediction of first pixel to reduce the amount of information of the prediction error, then the embedding process and chaos image content, serious damage image quality, and achieve the purpose of encryption. The local linear prediction module, prediction the effect of precise, making a large number of embedded information can be.2. for the existing method of embedding watermark cloud scene defect hiding algorithm the cloud cipher domain image transform based on reversible reversible information hiding algorithm, propose a cloud cipher domain and user terminal independent reversible information ", has practical application value. The method is based in a reversible image transform, the secret image will be protected by another plaintext plaintext image as reference for transformation, and get target image Let's express similar image transform, image semantic content covered secret protection target image content. Because the transformed image is clear, so the cloud server can embed watermarking information hiding algorithm is an arbitrary plaintext reversible information, the algorithm is independent of the user visible watermarking algorithm end of sender and receiver.3. cloud ciphertext domain reversible video watermarking embedding for cloud scene, proposed a novel video watermarking algorithm in cipher domain reversible. The reversible visible watermark algorithms are based on the plain image, to achieve reversibility by recording the original image and watermark image. The difference between the carrying secret because no correlation between the cipher image pixels, resulting in prediction the error is still a large number of random distribution, the difference between to be reversibly embedded into the image. In this paper, it is in the original image Some space is reserved in the space. After encrypting, the space is used to embed visible watermark, so as to avoid the record of the difference. The reserved pixels in the reserved space can be embedded in the non watermark region by the plaintext reversible information hiding method, so that it can be reconstructed nondestructively.

【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP309.7

【相似文献】