摘要
通过提取如图像的颜色、纹理、形状和空间关系等底层特征来建立索引,然后通过计算比较这些特征和查询条件之间的相似度来判断2个图像的相似程度,将相似度较高的图像显示给用户,进而完成图像检索.通过以建立医学图像的特征矢量作为其索引,使用医学图像的多维特征进行相似查询,可提高临床医师的确诊率.
Methods： The technology of content-based image retrieval in medicine is to inquire the related image or information based on the content of the given image. First it＇ s image pretreatment from optics and digital medicine angles. Second the image character extracting is drawn in, discussing the super ion of each method of extracting, and finally the main method is decided. Results： The method of content-based image matching is to analyze the image information ground on the low level character of the medical image,accord- ing to color, grain, shape etc. Conclusion： Through building up the image character index, the matching method based on the multimensional characters to inquire makes sense in raising the rate of making a deftnite diagnosis, and promoting the informatization of the autonomous region.
出处
《重庆工学院学报（自然科学版）》
2008年第8期139-143,共5页
Journal of Chongqing Institute of Technology
基金
新疆维吾尔自治区科学技术厅新疆少数民族科技骨干人才特殊培养科研专项基金资助项目
关键词
图像处理
特征提取
图像分割
纹理分析
形状分析
image processing
feature extraction
image segmentation
texture analysis
shape analysis
