【作者】 黄晓晴；

【导师】 于盛林；

【作者基本信息】 南京航空航天大学 ， 测试计量技术及仪器， 2010， 博士

【摘要】 在诸多图像压缩方法中,分形图像压缩方法是当今图像压缩领域中较新的方法之一。图像分形编码以其新颖的编码思想,较高的压缩比,得到了人们的普遍关注。而随着图像分形编码理论的深入研究,分形编码在其他领域的应用也日趋广泛。其中基于分形编码的数字水印技术引起了人们的极大关注,出现了多种分形编码数字水印算法。本文在研究分形编码的基础上,研究基于分形编码的数字水印算法,针对现有分形编码数字水印技术存在的问题,提出了三种不同特点的分形编码数字水印算法。同时,在分析分形编码参数特征的基础上,研究其在图像检索中的应用价值,提出了一种基于分形编码参数核密度估计的图像检索算法。本文首先介绍了图像基本分形编码算法,描述了分形编码的数学基础、编码方法和步骤。同时给出了基本分形编码的仿真实验结果。简要介绍了数字水印技术以及现有的基于分形编码的数字水印技术,总结了现有分形编码数字水印技术的优缺点,为后续分形编码数字水印技术的研究指明了方向。接着,研究了正交化分形编码算法,证明了正交化分形编码参数在迭代过程中的不变性,找出在分形编码参数中可以嵌入水印信息的有效途径。并且,本文通过实验验证了正交化分形解码可以快速收敛,使得解码速度得以提高。在研究正交化分形编码的基础上,依据自适应乘性嵌入准则,构造水印嵌入变换函数,提出了一种基于分形编码的灰度水印嵌入算法。该算法弥补了传统的分形编码数字水印技术只能将0、1序列作为水印嵌入的不足,首次给出了在分形编码参数中嵌入灰度水印的算法。分析了分数阶微积分伪随机序列的特点,并利用此序列对水印加密,提高了水印安全性。实验结果和数据表明,算法可行,且较之传统方法具有更好的鲁棒性。随后,提出了一种基于分形编码的盲水印算法。算法在保证可以嵌入灰度水印的基础上,实现了盲提取。算法基于参数量化调制,通过构造水印变换函数,实现了水印的嵌入。同时,分析了分形编码中拼贴误差的特性,并将此特性应用到水印嵌入中,提出了基于拼贴误差分类的嵌入强度调节方法,使得水印的鲁棒性和不可感知性有了明显提高。以图像认证为目的,提出了一种基于分形编码的脆弱性盲水印算法。该算法将认证水印嵌入分形变换域,又将分形编码参数作为水印嵌入到宿主图像中,使算法在篡改定位的同时,还可以自动恢复原图像。同时,本文还提出了二次水印嵌入策略,降低了图像认证的虚警率和漏检率。实验数据表明,算法可行有效,并且具有良好的篡改定位和恢复原图像的能力。利用核密度估计方法,对分形编码参数的统计特征做出分析,构造检索索引,提出了一种基于分形编码参数核密度估计的图像检索算法。与现有常用的直方图方法相比,该方法进一步提高了检索准确率及检索速度。最后,对全文研究做出总结,就分形编码数字水印技术研究中一些尚待解决的问题进行了讨论,对分形编码在数字水印及其他领域的应用研究做出展望。更多还原

【Abstract】 Fractal image compression, as one of the most popular techniques in the image compression fields, has novel decoding method, higher compression ratio. However, along with the deeply research on fractal image compression, the application of fractal coding in many other fields has become more and more widely. Especially in digital watermarking field, many different methods using fractal image coding appeared.In this paper, to solve the problems of the existing fractal image watermarking, three novel watermarking methods are proposed. Furthermore, based on the investigation of fractal image coding parameters, this paper also proposes a new image retrieval method using kernel density estimation of fractal coding parameters.The basic fractal coding method is reviewed. The mathematical basis of fractal coding is described, and the coding process, experimental results are proposed. Digital watermarking techniques, especially based on fractal coding techniques are introduced. For the subsequent fractal coding watermarking research, this paper summarizes the advantages and disadvantages of the existing techniques.The orthogonal fractal coding method is analyzed. Also, to get the effective way for embedding watermarking into fractal coding parameters, it has been proved that one of the parameters would be unchanged in the iterative processing. Experimental results also show that the orthogonal fractal coding method can converge very fast in image decoding.Based on the analysis of orthogonal fractal coding, according to the rules of adaptive multiplication embedding, a formula for watermarking embedding is proposed for embedding gray image as watermarking to the fractal coding image. Because traditional methods can only use 0, 1 sequence as the watermarking, not gray image, so, it is the compensation for the traditional methods.Experimental results also show that this method is feasible, and robust against many attacks. Another blind watermarking method based on fractal coding is proposed. This method can extract watermarking blindly, even if the watermarking is gray image. Based on the quantitative modulation, a formula for watermarking transform is proposed. On the other hand, the collage error of the fractal coding parameters is analyzed. Applied this characteristics into the proposed method, watermarking is much more robust and imperceptible.For image authentication, a fragile digital watermarking method is proposed. This method embeds watermarking into fractal transform, and also embeds fractal coding parameters as watermark into host image. It can not only locate altered areas but also automatically recover the original image through extracting the fractal coding parameters. Also, to reduce the false alarm rate and missed rate, the way for embedding watermarking secondly is proposed. Experimental results show that the proposed method is capable of tamper localization, and effective of automatic recovery.A statistical method, called kernel density estimation, is used for analyzing fractal coding parameters. And then, the fractal signatures are extracted for texture image retrieval. Experimental results show that this method has not only higher retrieval rate but also faster retrieval speed than the existing methods.A brief conclusion is made at the end of this paper. The unresolved problems under the application of fractal coding in watermarking are discussed. The outlook of fractal coding application in digital watermarking and other fields is presented.更多还原