Image quality comparison between JPEG and JPEG2000. II. Scene dependency, scene analysis, and classification
Image quality comparison between JPEG and JPEG2000. II. Scene dependency, scene analysis, and classification
复制标题
DOI：
10.2352/j.imagingsci.technol.(2007)51:3(259
复制
发表时间：
2007-05
影响因子：
1
通讯作者：
S. Triantaphillidou;E. Allen;R. E. Jacobson
中科院分区：
文献类型：
--
作者：
S. Triantaphillidou;E. Allen;R. E. Jacobson
图像质量评估表明 JPEG 和 JPEG2000 压缩方案都依赖于场景内容。本文解决了图像质量评估中的场景依赖性和场景敏感性问题，并提出图像分析作为根据人类观察者在判断图像质量时参考的基本固有场景属性对测试场景进行分组的手段。进行实验工作是为了研究场景内容与从[E. Allen、S. Triantaphillidou 和 R. E. Jacobson，“JPEG 和 JPEG2000 之间的图像质量比较。l. 心理物理学调查”，J. Imaging Sci。技术。 51, 248 (2007)]。本工作中使用的测试图像的内容是使用简单的图像分析措施进行分析的，这些措施量化了各种图像特征，例如原始场景对比度和全局亮度、主导线的数量、场景繁忙度（此处定义为指示存在或不存在细节的场景/图像属性）以及场景内的平坦区域。获得了初步结果和结论，并提出了建议，为进一步研究图像质量测量的场景依赖性和场景分类奠定了基础。
Image quality assessments have shown that both JPEG and JPEG2000 compression schemes are dependent on scene content. This paper addresses the problem of scene dependency and scene susceptibility in image quality assessments and proposes image analysis as a means to group test scenes, according to basic inherent scene properties that human observers refer to when they judge the quality of images. Experimental work is carried out to investigate the relationship between scene content and the subjective results obtained from experimental work carried out in [E. Allen, S. Triantaphillidou, and R. E. Jacobson, "Image quality comparison between JPEG and JPEG2000. l. Psychophysical Investigation", J. Imaging Sci. Technol. 51, 248 (2007)]. The content of the test images used in this work is analyzed using simple image analysis measures that quantify various image features, such as original scene contrast and global brightness, amount of dominant lines, scene busyness (defined here as a scene/image property indicating the presence or absence of detail), and flat areas within the scene. Preliminary results and conclusions are obtained and suggestions are made to form a basis for further studies on scene dependency and scene classification with respect to image quality measurements.