基于商空间理论的海量生物信息多粒度表示方法  

Multi-granularity modeling for the massive biological information in micro manipulation based on the quotient space theory

作　　者：贺丽梅[1] 尹昕[1] 刘晓倩[1] 赵智伟[1] 

机构地区：[1]天津师范大学电子与通信工程学院,天津300387

出　　处：《天津师范大学学报（自然科学版）》2016年第3期28-32,共5页Journal of Tianjin Normal University：Natural Science Edition

基　　金：天津市自然科学基金资助项目(13JCYBJC15800);天津师范大学大学生创新创业训练计划资助项目(201553)

摘　　要：微操作中,为研究微小目标的生物机理,需要连续获取海量微小目标的生物信息并进行特征分析与建模.针对微操作中海量生物信息多尺度、多层次和结构复杂的特点,利用商空间理论实现海量生物信息的多粒度表示,并将提取的多粒度特征运用于对显微图像进行匹配,以验证本研究算法的有效性.结果表明:算法全面且有效地描述了显微图像目标灰度、纹理和尺度空间方面的特征信息,可以为图像匹配、图像识别和图像检索等图像处理过程提供自适应特征信息,且满足实时性要求.In micro manipulation, massive tiny biological information need continuously obtaining and its characteristic analysis and modeling need taking to study the mechanism of micro biological objects. For the features like multi-scale, multi-level and complex structure of massive biological information, an algorithm of multi-granularity modeling for massive biological information based on the quotient space theory was realized, and the extracted multi-granularity features were used to microscopic image matching to verify the effectiveness of the algorithm. The resuhs show that multi-granularity fea- tures like gray features, texture features and scale space features are described effectively and comprehensively using this algorithm, and the algorithm provides adaptive features and satisfy the request of real time operation for image processing such as image matching, image recognition and image retrieval.

关 键 词：微操作 显微图像特征提取 商空间理论 粒度计算 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]