摘要
本文提出一种基于Fourier描绘子匹配脑部切片的算法 ,用于汇聚组合来自不同动物的生理数据。首先 ,将大鼠中脑组织切片数字化 ,用三条封闭特征轮廓表示 ,一组大鼠中脑 38张的连续切片被作为与其它任何切片匹配的标准模板。然后 ,将每一切片的轮廓作Fourier展开 ,通过计算两切片对应轮廓的尤拉距离 ,结合三轮廓的权重 ,定义两切片间的“相似函数”。最后 ,根据实验组切片与标准模板切片之间相似函数的大小实现两者的匹配。本算法已成功应用于调幅声敏感听觉细胞功能组织的研究。
We developed an algorithm based on Fourier Descriptors(FD) to match brain sections and specially aimed to pool electro physiological data from individual animals and composite the results on a single brain. To test this algorithm, curvilinear features were first extracted from each section image of the midbrain of the rat resulting in three closed boundaries. A set of boundaries of 38 consecutive sections was then taken as standard templates for matching with any other brain sections processed in a similar way. Each section was represented by Fourier descriptors of these boundaries. Matching was done by minimizing the similarity function between an experimental brain section and the set of standard templates. This algorithm has been successfully used for the study of functional organization of AM sound sensitive auditory neurons in the midbrain.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2000年第4期437-439,共3页
Journal of Biomedical Engineering
