摘要
针对医学图像数据库,提出一种新的基于纹理和形状的图像检索方法。非下采样Contourlet变换具有多尺度多分辨率分析和平移不变性的特点,用来提取纹理特征。Zernike矩作为一种基于区域的形状描述子,具有良好的旋转不变性,用来提取形状特征.然后综合两种特征。对CT图像数据库进行检索实验,实验结果表明,该方法具有良好的检索性能,并具有平移、尺度、旋转不变性。
Aiming at medical image database, a novel medical image retrieval algorithm based on texture and shape information is proposed. The nonsubsampled Contourlet transform provides a flexible multi-scale, multi-direction and shift-invariant image representation. It is used for texture feature extraction. As a region-based shape descriptor, Zernike moments have good rotation invariance. It is used to extract shape feature. Then, the two features are integrated into image retrieval. A CT image database was retrieved using this method. Experimental results show that the new method has good retrieval performance and the properties of translation, scaling and rotation invariance.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2009年第6期1275-1280,共6页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(60274099)资助项目
