摘要
针对医学图像全局特征检索不能很好地体现局部特征,而基于分割后各区域特征的检索又过分依赖于复杂的图像分割算法,提出了环形划分图像的分割方法.在基于环形分割的条件下,分别提取医学图像的颜色和纹理等内容特征,并在此基础上,实现医学图像检索.试验表明,基于环形分割的医学图像检索能够较好地体现图像所记录的颜色空间信息,而且具有旋转偏移不变性的特点,大大提高了医学图像检索的正确率与查全率.
In view of some disadvantages in present medical image retrieval, this article presents a new method based on annular segementation. Under such circumstance, the authors collected some samples of medical images and analysed the features of their colors and textures, and then established the image retrieval system. The experiment indicated that the medical image retrieval, based on annular segmentation, can record the information about color space, rotation invariability and excursion invariability. Thus, the accuracy of image retrieval has been greatly improved.
出处
《江苏大学学报（自然科学版）》
EI
CAS
北大核心
2008年第2期155-159,共5页
Journal of Jiangsu University：Natural Science Edition
基金
国家自然科学基金资助项目(60572112)
江苏省软件与集成电路专项基金资助项目([2005]196)
关键词
医学图像
环形分割
方形分割
图像检索
旋转偏转
medical image
annular segmentation
square segmentation
image retrieval
rotating and excursion
