基于小波与分形的图像压缩-计算机应用技术专业论文.docxVIP

中文摘要 中文摘要 I万方数据 I 万方数据 作为分析工具的小波变换和作为几何语言的分形理论相结合在图像压缩中的应 用已经成为近来众多学者的研究热点，但它还不成熟，还存在一些问题。如小波变换 后因截短的能量损失较大、分解后高频的能量分布还较多、分形编码由于寻找最优匹 配块需要较大的编码时间以及编码与解码的不对称等等。 针对上述问题，本文进行一些研究，取得如下成果： （1）为减少信号做卷积后因截短产生的能量损失和小波基选取对图像变换高频 能量分布，在第三章给出了顺序组合最优匹配小波的构造，以及它在图像中的应用和 效果。 （2）为了减少分形编码的时间，依据人眼对图像非线性的感知及图像本身分块 后平坦块和尖锐块数量分布不均的特点，给出了一种基于人眼视觉特性的感兴趣区分 形图像压缩编码方法和一种基于均方差的分类—邻域分形编码方法。把图像分成感兴 趣区和背景区，分别采用不同大小的块来进行分形编码，以及利用块的均方差特点来 搜索匹配块。 （3）传统基于 Banach 不动点定理的编码与解码是不对称的，为了克服这种缺点， 文中给出了可控变参数不动点定理，以及基于它的分形解码方法。改进后的解码方法 能更有效的控制解码过程，较好的解决了不对称问题。原来 Banach 不动点定理解码 方法只是本文方法的一种特殊情况。 （4）综合了以上原理和方法，给出一种改进的基于小波变换的分形图像压缩方 法。从实验的结果来看，改进后的方法在解码图像质量和解码时间都有较好的效果。 关键词：图像压 HYPERLINK \l _bookmark1 1缩；小波变换；分形编码；感兴趣区；分类—邻域；可控渐进。 注：文中的图表采用画图板制作。 Abstract Abstract II万方数据 II 万方数据 The application of the wavelet transformation of analyzing tool and the fractal theory of geometry language in the image compression has become the hot spot of multitudinous scholars researching. However, it is imperfect and has some problems, such as more energy loses in the wavelet transformation, more high-frequency energy distributions after the decomposition, more time cost during seeking the most superior code block, more asymmetry of the coding and the decoding and so on. In order to solve the problems above, some results has been proposed in this paper, and the main points are: In order to decrease energy loses in the wavelet transformation and high-frequency energy distributions after the decomposition to the greatest extent, a kind of the most superior match wavelet by the order is constructed and its application is proposed in chapter 3. In order to reduce the time cost during seeking the most superior code block, the methods of Region of Interest fractal image compression and class and neighborhood based on mean-variance fractal image compression are proposed. Because the sensation of human eyes is non-linear and the amount of the smooth block and the sharp block after image blocking is non-uniformly distributed, the image is