【作者】 周志勇；

【导师】 周激流；

【作者基本信息】 四川大学 ， 通信与信息系统， 2004， 硕士

【摘要】 静止图像压缩技术一直是信号处理方面的一个重要课题，因为随着信息技术的发展以及应用的扩展，人们对信息量的要求越来越大，特别是随着计算机处理能力的提高和网络技术的发展，人们都期望有更强的压缩技术出现，这种技术除了能够在有限的空间和带宽资源下存储与传播更多的图像数据外，还要求具有许多新的特性，如渐进编解码、ROI等，同时算法还应具有更广泛的可适用范围，如包括网络、医学、遥感等各种图像应用领域。 在过去的三十年里，JPEG专家组一直是这个领域标准的主要制定者，而JPEG标准也是上个世纪使用最广泛的标准之一。但随着技术的发展和应用的需要，JPEG已经不能适应现代网络以及信息处理技术上的要求，因此，JPEG专家组在2000年制定了新一代图像压缩标准系列——JPEG-2000。随着新一代图像压缩标准——JPEG-2000的制定，静态图像编码技术进入了一个新的时代，利用第二代小波变换以及丰富的编码技术，图像的压缩效率不仅得到了进一步的提高，更重要的是，新的标准具有更大的灵活性和更多的选择，从而适应了以上提到的趋势和要求；另外，新标准系列不仅制定了静止图像的压缩标准，还包括了运动图像、动画等其它一系列与图像相关的标准，应该说，标准基本涵盖了各种图像(如二进制、彩色图等)、各种应用方式(如运动图像、动画等)和各个应用领域(如医学、网络等)。 介于JPEG-2000标准的重要性以及本身所包含的技术对于现代通信以及图像压缩理论的影响，包括其思想对其它应用领域所起到的借鉴作用，本论文对该标准进行了深入的讨论，特别是对其核心技术——小波变换和编码部摘要分—做了重点的探讨:针对算法在时间上所存在的缺陷，我认为主要是时间和复杂度上的缺陷，当然也包括码流组织中存在的一些效率问题，我们利用现代小波变换系数编码中刚刚提出的SPECK算法，结合比特平面编码，提出了一种新的算法，新算法在保持较高压缩率的情况下，在编程、运算复杂度、运行时间上都要有效得多，同时该算法相对过去的压缩标准还具有一定的优势;另外，针对JPEG一2000在应用上所蕴涵的价值，我们对其在运动图像中的应用进行了分析，并给出了一定的结论。 本文主要可以分为三部分。第一部分主要对图像压缩编码理论基础作一简单的介绍，主要是编码技术的发展以及相关基础理论;第二部分重点介绍JPEG一2000标准算法及其相关技术，包括小波变换，EBCOT编码以及算法中包含的多种特性及其实现原理，另外对该算法进行了一些分析、比较和测试，说明了该算法相比其它算法的优点;第三部分则介绍了我们对编码过程进行的改进以及一些设想，主要包括SBCA算法以及JPEG一2000在运动图像(也可以是视频图像)中的一个应用。关键词:JPEG一2000小波变换提升算法EBCOT(优化截取的嵌入式编码) sBeA(基于sPEeK的编码算法)更多还原

【Abstract】 Still image coding has always been the most important thesis in the signal process area. Because with development of technology and application, people have much more and more demand on information quantity, especially with the boost of computer’s calculation power and the development of networks, people has been longing for a high-tech coding technology. The new technology not only can save and transmit more image data under the constraint space and bandwidth, but also must has some extra functions, such as progressive transmission by pixel accuracy and resolution, ROI, etc, on the same time, this new technology must also be able to applied in more area as board as possible, for example, networks, medicine, remote control.In the past thirty years, JPEG group has always been the standard establisher in the still image coding area and JPEG is also one of the most popular standard in the last century. But with the development of new technology and the demand of application, JPEG can’t adapt to the demand of networks and signal processingarea, for this, JPEG group make out the new still image coding standard-JPEG-2000. With the new still image coding standard JPEG-2000 being confirmed, still image coding technology step into a new era. Using the second-class wavelet transformation and rich coding technologies, we not only can increase compression rate, but more importantly, the new standard has more flexibility and choices, which accommodated the trend. In addition, the new standard not only include still image coding standard but also include a series of standard which relate to motion image, carton etc. In one word, new standard include almost every image type (binary, color image), every image application (motion, carton etc) and every application areas (such as medicine, networks).Due to the importance of JPEG-2000 standard and its effect on the modern communication and image coding. The paper will discuss it thoroughly, especially pay attention to its coding technology; According to its flaws in time consuming and complexity, we propose a new coding algorithm base on the SPECK and bit-plane coding, which not only have low complexity and time consuming but also easy to programme. In allusion to the its application potential, we also want to ameliorate the standard and expand it’s utility area.The paper can be divided into three parts. In the first part, we introduce the basic theory of image coding briefly, which mainly devote to the development of coding technology and it’s relate basic theory. In second part, we will discuss JPEG-2000 standard and it’s relate technologies in detail, including wavelet transformation, EBCOT and some test result. The last part will pay attention to our amelioration and some proposes, mainly about SBCA algorithm and the application of the JPEG-2000 in motion image.更多还原