【作者】 徐鲁安；

【导师】 叶懋冬；

【作者基本信息】 浙江大学 ， 应用数学， 2003， 硕士

【摘要】 对图象质量评价的研究已成为图象信息工程中重要的基础研究课题之一。它不仅有重要的理论价值，而且有广泛的应用背景，使其得到越来越多的关注。特别是图象压缩技术的迅速发展大大促进了对图象质量评价方法的研究。 文章介绍了“图像质量”的概念，并以人眼的视觉特性为出发点，基于小波理论提出了新的图象质量评价方法，并证明了该方法的有效性。 首先，文章明确了“图像质量”的概念，指出了“图像质量评价”的重要意义，并给出了图像质量评价方法发展的历史和分类。作者还通过分析多种有代表性的图像评价方法，对现有方法的特点和优劣进行了总结和比较。 文章的主体给出了一种新的基于小波变换的图象质量评价方法。在第三章首先从低级的视觉感觉和高级的视觉心理角度，深入地描述了人眼的视觉特性。并在第四章介绍了本文提出方法所基于的理论基础——小波变换，对其概念以及相关性质进行了阐述。第五章详细论述了基于小波变换的图象质量评价方法。该方法通过提取图象的亮度，清晰度，相似度，尺度失真度等多种指标来综合评价图象质量。与传统的质量评价标准相比，该方法对图象质量的评价更全面、更符合人眼的视觉感受。最后，通过实验证明了该方法的有效性。更多还原

【Abstract】 Researches on evaluation of digital image quality are meaningful and basic for the image processing engineering. It not only is essential on theory importance, but also has taken an extensive appllication background, which makes the evaluation be paid much attention to. Especially, the rapid development of image compressing technology accelerates researches on quality evaluation of digital images.The dissertation introduces the concept of " image quality " .From the point of view of people eyes’ visual characteristics, a new method to evaluate image quality based on the wavelet theory is puts forward and then its efficiency is proven,At first the author describes the concept of " image quality" detailedly, po i nts out the great sense of " image quality evaluation " . After the description of the development history and classification of the methods of image quality evaluation some typical methods is analyzed, their characteristics, good qualities and shortcomings are compared and summarized .The main goal of this dissertation is to present a new method of image quality evaluation based on wavelet transform. In Chap 3 people eyes’ visual characteristics are described accurately and deeply from low-class visual feeling to high-class visual psychology. Wavelet transform, the theoretic base of the new method is introduced in Chap 4 with its concept and some properties explained. In Chap 5 the dissertation discusses the new method of image quality evaluation based on wavelet transform. The method describes and evaluates the quality of image through wavelet transform extracting luminance, definition and corelation. Compared with the traditional method it’s evaluation effect will be more comprehensive and consistent with human perception. In the end, the experimental results prove that the method is accurate and efficient.更多还原