基于DICOM文本与内容的医学图像检索系统研究

摘要

为辅助医师从海量医学图像中快速查找所需图像,提出了一种将DICOM文本信息和图像底层特征相结合的医学图像检索方法.从DICOM图像的文件头中提取文本信息,进行图像预筛选,提取图像的纹理及形状特征,通过设定权重将这两种特征相结合,进行基于内容的图像检索.实验结果表明,综合纹理和形状特征的检索方法优于单一特征检索方法,联合DICOM文本信息和底层特征的方法能明显提高检索的效能,使检索结果更接近于人的视觉理解.%To assist physicians to quickly find the images from mass medical images, an approach is presented, which combines header text information and iow-level features of digital imaging and communications in medicine (DICOM) images to perform retrieval on medical image databases.Firstly, thc text information is extracted from DICOM header for the pre-filtering of the images, and then texture and shape features arc extracted and combined with different weights to retrieve similar images.The results of experiments show that the proposed method outperformed that based on single kind of features.By combining the DICOM header information with lowlevel content features, the performance of medical image retrieval is improved.