摘要
为了克服传统方法在检测图像色偏上的局限性问题,提出了一种基于图像分析的图像色偏检测方法.并且在该方法的基础上,给出了一种基于图像子样的两步判别方法,既有效提高了算法的运行效率,又通过消除图像中主色调对色偏检测的影响,提高了检测的准确率和可靠性.依据所设计的评价试验,对大量不同主题的图像进行了评测.试验表明,文中所提出的方法与主观评测结果具有较好的吻合.在此基础上,成为一个数字成像系统图像品质评价工具,并在工程实际中得到了应用.
In order to overcome the limit of traditional methods of evaluating digital image color errors, a new approach based on image analysis is presented. Meanwhile, the technique of evaluating color errors has been implemented with the approach to enhance the effectiveness, reliability and accuracy of the evaluation. The technique is divided into two phases. Both phases are carried out on a sub-sampled version of the image. And the dominant colors are eliminated in the second phase. On this basis, an experiment has been designed with many images. The comparison of the evaluating results of this technique to that of human shows that the approach proposed is accurate and reliable. An image quality evaluating tool is implemented based on this approach, and has been applied in some factories.
出处
《江苏大学学报（自然科学版）》
EI
CAS
2004年第5期430-433,共4页
Journal of Jiangsu University：Natural Science Edition
基金
国家自然科学基金资助项目(60273040)
关键词
数字图像
色偏
图像分析
图像子样
digital image
color error
image analysis
sub-ampled image
